ci(cli): change release workflow to manual dispatch from main

- Trigger via Actions UI with a version input (e.g. "0.1.0")
- Only runs on main branch
- Creates git tag cli/v<version> automatically
- Then GoReleaser builds all 6 binaries and creates the GitHub Release

.gitattributes

@@ -9,6 +9,4 @@ packages/browseros/chromium_patches/**/*.py linguist-generated
 scripts/*.py linguist-generated
 # Mark build directories as generated
 build/* linguist-generated
-# Mark eval/test framework as vendored so it's excluded from language stats
-packages/browseros-agent/apps/eval/** linguist-vendored
 docs/videos/** filter=lfs diff=lfs merge=lfs -text

.github/workflows/pr-title.yml

@@ -2,7 +2,7 @@ name: PR Conventional Commit Validation
 
 on:
   pull_request:
-    types: [opened, edited]
+    types: [opened, synchronize, reopened, edited]
 
 permissions:
   pull-requests: write

.github/workflows/release-agent-extension.yml

@@ -1,148 +0,0 @@
-name: Release BrowserOS Extension
-
-on:
-  workflow_dispatch:
-
-concurrency:
-  group: release-agent-extension
-  cancel-in-progress: false
-
-jobs:
-  release:
-    if: github.ref == 'refs/heads/main'
-    runs-on: ubuntu-latest
-    permissions:
-      contents: write
-      pull-requests: write
-    defaults:
-      run:
-        working-directory: packages/browseros-agent/apps/agent
-
-    steps:
-      - uses: actions/checkout@v6
-        with:
-          fetch-depth: 0
-
-      - uses: oven-sh/setup-bun@v2
-
-      - name: Install dependencies
-        run: bun ci
-        working-directory: packages/browseros-agent
-
-      - name: Build and zip extension
-        run: bun run codegen && bun run zip
-        env:
-          VITE_PUBLIC_BROWSEROS_API: https://api.browseros.com
-
-      - name: Get version and zip path
-        id: version
-        run: |
-          echo "version=$(node -p "require('./package.json').version")" >> "$GITHUB_OUTPUT"
-          echo "release_sha=$(git rev-parse HEAD)" >> "$GITHUB_OUTPUT"
-          ZIP_FILE=$(ls "$(pwd)/dist/"*-chrome.zip | head -n 1)
-          echo "zip_path=$ZIP_FILE" >> "$GITHUB_OUTPUT"
-          echo "zip_name=$(basename "$ZIP_FILE")" >> "$GITHUB_OUTPUT"
-
-      - name: Generate release notes
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          AGENT_PATH="packages/browseros-agent/apps/agent"
-          CURRENT_TAG="agent-extension-v${{ steps.version.outputs.version }}"
-          PREV_TAG=$(git tag -l "agent-extension-v*" --sort=-v:refname | grep -v "^${CURRENT_TAG}$" | head -n 1)
-
-          if [ -z "$PREV_TAG" ]; then
-            echo "Initial release" > /tmp/release-notes.md
-          else
-            COMMITS=$(git log "$PREV_TAG"..HEAD --pretty=format:"%H" -- "$AGENT_PATH")
-
-            if [ -z "$COMMITS" ]; then
-              echo "No notable changes." > /tmp/release-notes.md
-            else
-              echo "## What's Changed" > /tmp/release-notes.md
-              echo "" >> /tmp/release-notes.md
-
-              while IFS= read -r SHA; do
-                SUBJECT=$(git log -1 --pretty=format:"%s" "$SHA")
-                PR_NUM=$(gh api "/repos/${{ github.repository }}/commits/${SHA}/pulls" --jq '.[0].number // empty' 2>/dev/null)
-
-                # Skip PR number if already in the commit subject (squash merges include it)
-                if [ -n "$PR_NUM" ] && ! echo "$SUBJECT" | grep -qF "(#${PR_NUM})"; then
-                  echo "- ${SUBJECT} (#${PR_NUM})" >> /tmp/release-notes.md
-                else
-                  echo "- ${SUBJECT}" >> /tmp/release-notes.md
-                fi
-              done <<< "$COMMITS"
-            fi
-          fi
-        working-directory: ${{ github.workspace }}
-
-      - name: Create GitHub release
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          TAG="agent-extension-v${{ steps.version.outputs.version }}"
-          RELEASE_SHA="${{ steps.version.outputs.release_sha }}"
-          TITLE="BrowserOS Extension - v${{ steps.version.outputs.version }}"
-
-          if git rev-parse "$TAG" >/dev/null 2>&1; then
-            echo "Tag $TAG already exists, skipping tag creation"
-          else
-            git tag "$TAG" "$RELEASE_SHA"
-          fi
-
-          if git ls-remote --tags origin "$TAG" | grep -q "$TAG"; then
-            echo "Tag $TAG already on remote, skipping push"
-          else
-            git push origin "$TAG"
-          fi
-
-          if gh release view "$TAG" >/dev/null 2>&1; then
-            echo "Release $TAG already exists, updating"
-            gh release edit "$TAG" --title "$TITLE" --notes-file /tmp/release-notes.md
-            gh release upload "$TAG" "${{ steps.version.outputs.zip_path }}" --clobber
-          else
-            gh release create "$TAG" \
-              --title "$TITLE" \
-              --notes-file /tmp/release-notes.md \
-              "${{ steps.version.outputs.zip_path }}"
-          fi
-        working-directory: ${{ github.workspace }}
-
-      - name: Update CHANGELOG.md via PR
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          VERSION="${{ steps.version.outputs.version }}"
-          DATE=$(date -u +"%Y-%m-%d")
-          BRANCH="docs/agent-extension-changelog-v${VERSION}"
-          CHANGELOG="packages/browseros-agent/apps/agent/CHANGELOG.md"
-
-          git checkout main
-
-          {
-            head -n 1 "$CHANGELOG"
-            echo ""
-            echo "## v${VERSION} (${DATE})"
-            echo ""
-            cat /tmp/release-notes.md
-            echo ""
-            tail -n +2 "$CHANGELOG"
-          } > /tmp/new-changelog.md
-          mv /tmp/new-changelog.md "$CHANGELOG"
-
-          git config user.name "github-actions[bot]"
-          git config user.email "github-actions[bot]@users.noreply.github.com"
-          git checkout -b "$BRANCH"
-          git add "$CHANGELOG"
-          git commit -m "docs: update agent extension changelog for v${VERSION}"
-          git push origin "$BRANCH"
-
-          gh pr create \
-            --title "docs: update agent extension changelog for v${VERSION}" \
-            --body "Auto-generated changelog update for BrowserOS Extension v${VERSION}." \
-            --base main \
-            --head "$BRANCH"
-
-          gh pr merge "$BRANCH" --squash --auto || true
-        working-directory: ${{ github.workspace }}

.github/workflows/release-agent-sdk.yml

@@ -1,27 +1,18 @@
-name: Release BrowserOS Agent SDK
+name: Release Agent SDK
 
 on:
   workflow_dispatch:
 
-concurrency:
-  group: release-agent-sdk
-  cancel-in-progress: false
-
 jobs:
   publish:
     if: github.ref == 'refs/heads/main'
     runs-on: ubuntu-latest
-    permissions:
-      contents: write
-      pull-requests: write
     defaults:
       run:
         working-directory: packages/browseros-agent/packages/agent-sdk
 
     steps:
       - uses: actions/checkout@v6
-        with:
-          fetch-depth: 0
 
       - uses: oven-sh/setup-bun@v2
 
@@ -40,129 +31,7 @@ jobs:
       - name: Test
         run: bun test
 
-      - name: Get version
-        id: version
-        run: |
-          echo "version=$(node -p "require('./package.json').version")" >> "$GITHUB_OUTPUT"
-          echo "release_sha=$(git rev-parse HEAD)" >> "$GITHUB_OUTPUT"
-
-      - name: Generate release notes
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          SDK_PATH="packages/browseros-agent/packages/agent-sdk"
-          CURRENT_TAG="agent-sdk-v${{ steps.version.outputs.version }}"
-          # Find the previous tag, excluding the current version's tag
-          # (which may already exist from a prior failed run)
-          PREV_TAG=$(git tag -l "agent-sdk-v*" --sort=-v:refname | grep -v "^${CURRENT_TAG}$" | head -n 1)
-
-          if [ -z "$PREV_TAG" ]; then
-            echo "Initial release" > /tmp/release-notes.md
-          else
-            # Get commits scoped to the SDK directory
-            COMMITS=$(git log "$PREV_TAG"..HEAD --pretty=format:"%H" -- "$SDK_PATH")
-
-            if [ -z "$COMMITS" ]; then
-              echo "No notable changes." > /tmp/release-notes.md
-            else
-              echo "## What's Changed" > /tmp/release-notes.md
-              echo "" >> /tmp/release-notes.md
-
-              # For each commit, find the associated PR and format with author
-              CONTRIBUTORS=""
-              while IFS= read -r SHA; do
-                # Get commit subject and author
-                SUBJECT=$(git log -1 --pretty=format:"%s" "$SHA")
-                AUTHOR=$(git log -1 --pretty=format:"%an" "$SHA")
-                GITHUB_USER=$(gh api "/repos/${{ github.repository }}/commits/${SHA}" --jq '.author.login // empty' 2>/dev/null)
-
-                # Find associated PR number
-                PR_NUM=$(gh api "/repos/${{ github.repository }}/commits/${SHA}/pulls" --jq '.[0].number // empty' 2>/dev/null)
-
-                # Format line: skip PR number if already in the commit subject
-                # (squash merges include "(#123)" in the subject automatically)
-                if [ -n "$PR_NUM" ] && ! echo "$SUBJECT" | grep -qF "(#${PR_NUM})"; then
-                  echo "- ${SUBJECT} (#${PR_NUM})" >> /tmp/release-notes.md
-                else
-                  echo "- ${SUBJECT}" >> /tmp/release-notes.md
-                fi
-              done <<< "$COMMITS"
-            fi
-          fi
-        working-directory: ${{ github.workspace }}
-
       - name: Publish
         run: npm publish --access public
         env:
           NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}
-
-      - name: Create GitHub release
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          TAG="agent-sdk-v${{ steps.version.outputs.version }}"
-          RELEASE_SHA="${{ steps.version.outputs.release_sha }}"
-          TITLE="BrowserOS Agent SDK - v${{ steps.version.outputs.version }}"
-
-          # Create or reuse tag (idempotent for re-runs)
-          if git rev-parse "$TAG" >/dev/null 2>&1; then
-            echo "Tag $TAG already exists, skipping tag creation"
-          else
-            git tag "$TAG" "$RELEASE_SHA"
-          fi
-
-          # Push tag (skip if already on remote)
-          if git ls-remote --tags origin "$TAG" | grep -q "$TAG"; then
-            echo "Tag $TAG already on remote, skipping push"
-          else
-            git push origin "$TAG"
-          fi
-
-          # Create or update release
-          if gh release view "$TAG" >/dev/null 2>&1; then
-            echo "Release $TAG already exists, updating"
-            gh release edit "$TAG" --title "$TITLE" --notes-file /tmp/release-notes.md
-          else
-            gh release create "$TAG" --title "$TITLE" --notes-file /tmp/release-notes.md
-          fi
-        working-directory: ${{ github.workspace }}
-
-      - name: Update CHANGELOG.md via PR
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          VERSION="${{ steps.version.outputs.version }}"
-          DATE=$(date -u +"%Y-%m-%d")
-          BRANCH="docs/agent-sdk-changelog-v${VERSION}"
-          CHANGELOG="packages/browseros-agent/packages/agent-sdk/CHANGELOG.md"
-
-          # Return to main before branching
-          git checkout main
-
-          # Use head/tail to safely insert without sed quoting issues
-          {
-            head -n 1 "$CHANGELOG"
-            echo ""
-            echo "## v${VERSION} (${DATE})"
-            echo ""
-            cat /tmp/release-notes.md
-            echo ""
-            tail -n +2 "$CHANGELOG"
-          } > /tmp/new-changelog.md
-          mv /tmp/new-changelog.md "$CHANGELOG"
-
-          git config user.name "github-actions[bot]"
-          git config user.email "github-actions[bot]@users.noreply.github.com"
-          git checkout -b "$BRANCH"
-          git add "$CHANGELOG"
-          git commit -m "docs: update agent-sdk changelog for v${VERSION}"
-          git push origin "$BRANCH"
-
-          gh pr create \
-            --title "docs: update agent-sdk changelog for v${VERSION}" \
-            --body "Auto-generated changelog update for BrowserOS Agent SDK v${VERSION}." \
-            --base main \
-            --head "$BRANCH"
-
-          gh pr merge "$BRANCH" --squash --auto || true
-        working-directory: ${{ github.workspace }}

.github/workflows/release-cli.yml

@@ -1,4 +1,4 @@
-name: Release BrowserOS CLI
+name: Release CLI
 
 on:
   workflow_dispatch:
@@ -8,18 +8,13 @@ on:
         required: true
         type: string
 
-concurrency:
-  group: release-cli
-  cancel-in-progress: false
+permissions:
+  contents: write
 
 jobs:
   release:
     if: github.ref == 'refs/heads/main'
     runs-on: ubuntu-latest
-    environment: release-core
-    permissions:
-      contents: write
-      pull-requests: write
     defaults:
       run:
         working-directory: packages/browseros-agent/apps/cli
@@ -33,129 +28,23 @@ jobs:
         with:
           go-version-file: packages/browseros-agent/apps/cli/go.mod
 
-      - uses: oven-sh/setup-bun@v2
-        with:
-          bun-version: "1.3.6"
-
       - name: Run tests
-        run: make test
+        run: go test ./... -v
 
       - name: Run vet
-        run: make vet
+        run: go vet ./...
 
-      - name: Build all platforms
-        run: make release VERSION=${{ inputs.version }} POSTHOG_API_KEY=${{ secrets.POSTHOG_API_KEY }}
-
-      - name: Install dependencies
-        run: bun install
-        working-directory: packages/browseros-agent
-
-      - name: Upload to CDN
-        env:
-          R2_ACCOUNT_ID: ${{ secrets.R2_ACCOUNT_ID }}
-          R2_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
-          R2_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
-          R2_BUCKET: ${{ secrets.R2_BUCKET }}
-          R2_UPLOAD_PREFIX: cli
-          CLI_VERSION: ${{ inputs.version }}
+      - name: Create tag
         run: |
-          bun scripts/build/cli.ts \
-            --release \
-            --version "$CLI_VERSION" \
-            --binaries-dir apps/cli/dist
-        working-directory: packages/browseros-agent
-
-      - name: Generate release notes
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          CLI_PATH="packages/browseros-agent/apps/cli"
-          TAG="browseros-cli-v${{ inputs.version }}"
-          CHANGELOG_FILE="/tmp/release-changelog.md"
-          PREV_TAG=$(git tag -l "browseros-cli-v*" --sort=-v:refname | grep -v "^${TAG}$" | head -n 1)
-
-          if [ -z "$PREV_TAG" ]; then
-            echo "Initial release of browseros-cli." > "$CHANGELOG_FILE"
-          else
-            COMMITS=$(git log "$PREV_TAG"..HEAD --pretty=format:"%H" -- "$CLI_PATH")
-
-            if [ -z "$COMMITS" ]; then
-              echo "No notable changes." > "$CHANGELOG_FILE"
-            else
-              echo "## What's Changed" > "$CHANGELOG_FILE"
-              echo "" >> "$CHANGELOG_FILE"
-
-              while IFS= read -r SHA; do
-                SUBJECT=$(git log -1 --pretty=format:"%s" "$SHA")
-                PR_NUM=$(gh api "/repos/${{ github.repository }}/commits/${SHA}/pulls" --jq '.[0].number // empty' 2>/dev/null)
-
-                if [ -n "$PR_NUM" ] && ! echo "$SUBJECT" | grep -qF "(#${PR_NUM})"; then
-                  echo "- ${SUBJECT} (#${PR_NUM})" >> "$CHANGELOG_FILE"
-                else
-                  echo "- ${SUBJECT}" >> "$CHANGELOG_FILE"
-                fi
-              done <<< "$COMMITS"
-            fi
-          fi
-
-          cat "$CHANGELOG_FILE" > /tmp/release-notes.md
-          cat >> /tmp/release-notes.md <<'EOF'
-
-          ## Install `browseros-cli`
-
-          ### npm / npx
-
-          ```bash
-          npx browseros-cli --help
-          npm install -g browseros-cli
-          ```
-
-          ### macOS / Linux
-
-          ```bash
-          curl -fsSL https://cdn.browseros.com/cli/install.sh | bash
-          ```
-
-          ### Windows
-
-          ```powershell
-          irm https://cdn.browseros.com/cli/install.ps1 | iex
-          ```
-
-          After install, run `browseros-cli init` to point the CLI at your BrowserOS MCP server.
-          EOF
-        working-directory: ${{ github.workspace }}
-
-      - name: Create tag and release
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          TAG="browseros-cli-v${{ inputs.version }}"
           git config user.name "github-actions[bot]"
           git config user.email "github-actions[bot]@users.noreply.github.com"
+          git tag -a "cli/v${{ inputs.version }}" -m "browseros-cli v${{ inputs.version }}"
+          git push origin "cli/v${{ inputs.version }}"
 
-          if ! git rev-parse "$TAG" >/dev/null 2>&1; then
-            git tag -a "$TAG" -m "browseros-cli v${{ inputs.version }}"
-            git push origin "$TAG"
-          fi
-
-          CLI_DIST="packages/browseros-agent/apps/cli/dist"
-          gh release create "$TAG" \
-            --title "BrowserOS CLI - v${{ inputs.version }}" \
-            --notes-file /tmp/release-notes.md \
-            ${CLI_DIST}/*
-        working-directory: ${{ github.workspace }}
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
+      - uses: goreleaser/goreleaser-action@v6
         with:
-          node-version: "20"
-          registry-url: "https://registry.npmjs.org"
-
-      - name: Publish to npm
+          version: "~> v2"
+          args: release --clean
+          workdir: packages/browseros-agent/apps/cli
         env:
-          NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}
-        run: |
-          make npm-version VERSION=${{ inputs.version }}
-          cd npm
-          npm publish --access public
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/release-server.yml

@@ -1,147 +0,0 @@
-name: Release BrowserOS Server
-
-on:
-  workflow_dispatch:
-    inputs:
-      version:
-        description: "Release version (e.g. 0.0.80)"
-        required: true
-        type: string
-
-concurrency:
-  group: release-server
-  cancel-in-progress: false
-
-jobs:
-  release:
-    if: github.ref == 'refs/heads/main'
-    runs-on: ubuntu-latest
-    environment: release-core
-    permissions:
-      contents: write
-    defaults:
-      run:
-        working-directory: packages/browseros-agent
-
-    steps:
-      - uses: actions/checkout@v6
-        with:
-          fetch-depth: 0
-
-      - uses: oven-sh/setup-bun@v2
-        with:
-          bun-version: "1.3.6"
-
-      - name: Install dependencies
-        run: bun ci
-
-      - name: Prepare production env file
-        run: cp apps/server/.env.production.example apps/server/.env.production
-
-      - name: Validate version
-        id: version
-        env:
-          REQUESTED_VERSION: ${{ inputs.version }}
-        run: |
-          PACKAGE_VERSION=$(node -p "require('./apps/server/package.json').version")
-          echo "package_version=$PACKAGE_VERSION" >> "$GITHUB_OUTPUT"
-          echo "release_sha=$(git rev-parse HEAD)" >> "$GITHUB_OUTPUT"
-
-          if [ "$PACKAGE_VERSION" != "$REQUESTED_VERSION" ]; then
-            echo "Requested version $REQUESTED_VERSION does not match apps/server/package.json ($PACKAGE_VERSION)"
-            exit 1
-          fi
-
-      - name: Build release artifacts
-        run: bun run build:server:ci
-
-      - name: Verify release artifacts
-        run: |
-          mapfile -t ZIP_FILES < <(find dist/prod/server -maxdepth 1 -type f -name 'browseros-server-resources-*.zip' | sort)
-
-          if [ "${#ZIP_FILES[@]}" -eq 0 ]; then
-            echo "No server release zip files were produced"
-            exit 1
-          fi
-
-          printf 'Found release artifacts:\n%s\n' "${ZIP_FILES[@]}"
-
-      - name: Generate release notes
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          PACKAGE_VERSION: ${{ steps.version.outputs.package_version }}
-        run: |
-          SERVER_APP_PATH="packages/browseros-agent/apps/server"
-          SERVER_BUILD_DIR="packages/browseros-agent/scripts/build/server"
-          SERVER_BUILD_ENTRY="packages/browseros-agent/scripts/build/server.ts"
-          SERVER_RESOURCE_MANIFEST="packages/browseros-agent/scripts/build/config/server-prod-resources.json"
-          SERVER_WORKSPACE_PKG="packages/browseros-agent/package.json"
-          CURRENT_TAG="browseros-server-v$PACKAGE_VERSION"
-          PREV_TAG=$(git tag -l "browseros-server-v*" --sort=-v:refname | grep -v "^${CURRENT_TAG}$" | head -n 1)
-
-          if [ -z "$PREV_TAG" ]; then
-            echo "Initial release of browseros-server." > /tmp/release-notes.md
-          else
-            COMMITS=$(git log "$PREV_TAG"..HEAD --pretty=format:"%H" -- \
-              "$SERVER_APP_PATH" \
-              "$SERVER_BUILD_DIR" \
-              "$SERVER_BUILD_ENTRY" \
-              "$SERVER_RESOURCE_MANIFEST" \
-              "$SERVER_WORKSPACE_PKG")
-
-            if [ -z "$COMMITS" ]; then
-              echo "No notable changes." > /tmp/release-notes.md
-            else
-              echo "## What's Changed" > /tmp/release-notes.md
-              echo "" >> /tmp/release-notes.md
-
-              while IFS= read -r SHA; do
-                SUBJECT=$(git log -1 --pretty=format:"%s" "$SHA")
-                PR_NUM=$(gh api "/repos/${{ github.repository }}/commits/${SHA}/pulls" --jq '.[0].number // empty' 2>/dev/null)
-
-                if [ -n "$PR_NUM" ] && ! echo "$SUBJECT" | grep -qF "(#${PR_NUM})"; then
-                  echo "- ${SUBJECT} (#${PR_NUM})" >> /tmp/release-notes.md
-                else
-                  echo "- ${SUBJECT}" >> /tmp/release-notes.md
-                fi
-              done <<< "$COMMITS"
-            fi
-          fi
-        working-directory: ${{ github.workspace }}
-
-      - name: Create GitHub release
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          PACKAGE_VERSION: ${{ steps.version.outputs.package_version }}
-          RELEASE_SHA: ${{ steps.version.outputs.release_sha }}
-        run: |
-          TAG="browseros-server-v$PACKAGE_VERSION"
-          TITLE="BrowserOS Server - v$PACKAGE_VERSION"
-          mapfile -t ZIP_FILES < <(find packages/browseros-agent/dist/prod/server -maxdepth 1 -type f -name 'browseros-server-resources-*.zip' | sort)
-
-          git config user.name "github-actions[bot]"
-          git config user.email "github-actions[bot]@users.noreply.github.com"
-
-          if git rev-parse "$TAG" >/dev/null 2>&1; then
-            echo "Tag $TAG already exists, skipping tag creation"
-          else
-            git tag -a "$TAG" -m "browseros-server v$PACKAGE_VERSION" "$RELEASE_SHA"
-          fi
-
-          if git ls-remote --tags origin "$TAG" | grep -q "$TAG"; then
-            echo "Tag $TAG already on remote, skipping push"
-          else
-            git push origin "$TAG"
-          fi
-
-          if gh release view "$TAG" >/dev/null 2>&1; then
-            echo "Release $TAG already exists, updating"
-            gh release edit "$TAG" --title "$TITLE" --notes-file /tmp/release-notes.md
-            gh release upload "$TAG" "${ZIP_FILES[@]}" --clobber
-          else
-            gh release create "$TAG" \
-              --title "$TITLE" \
-              --notes-file /tmp/release-notes.md \
-              "${ZIP_FILES[@]}"
-          fi
-        working-directory: ${{ github.workspace }}

README.md

@@ -6,7 +6,6 @@
 [![Slack](https://img.shields.io/badge/Slack-Join%20us-4A154B?logo=slack&logoColor=white)](https://dub.sh/browserOS-slack)
 [![Twitter](https://img.shields.io/twitter/follow/browserOS_ai?style=social)](https://twitter.com/browseros_ai)
 [![License: AGPL v3](https://img.shields.io/badge/License-AGPL%20v3-blue.svg)](LICENSE)
-[![Docs](https://img.shields.io/badge/Docs-docs.browseros.com-blue)](https://docs.browseros.com)
 <br></br>
 <a href="https://files.browseros.com/download/BrowserOS.dmg">
   <img src="https://img.shields.io/badge/Download-macOS-black?style=flat&logo=apple&logoColor=white" alt="Download for macOS (beta)" />
@@ -23,183 +22,146 @@
 <br />
 </div>
 
-BrowserOS is an open-source Chromium fork that runs AI agents natively. **The privacy-first alternative to ChatGPT Atlas, Perplexity Comet, and Dia.**
+##
+🌐 BrowserOS is an open-source Chromium fork that runs AI agents natively. **The privacy-first alternative to ChatGPT Atlas, Perplexity Comet, and Dia.**
 
-Use your own API keys or run local models with Ollama. Your data never leaves your machine.
+🔒 Use your own API keys or run local models with Ollama. Your data never leaves your machine.
 
-> **[Documentation](https://docs.browseros.com)** · **[Discord](https://discord.gg/YKwjt5vuKr)** · **[Slack](https://dub.sh/browserOS-slack)** · **[Twitter](https://x.com/browserOS_ai)** · **[Feature Requests](https://github.com/browseros-ai/BrowserOS/issues/99)**
+💡 Join our [Discord](https://discord.gg/YKwjt5vuKr) or [Slack](https://dub.sh/browserOS-slack) and help us build! Have feature requests? [Suggest here](https://github.com/browseros-ai/BrowserOS/issues/99).
 
-## Quick Start
+## Quick start
 
-1. **Download and install** BrowserOS — [macOS](https://files.browseros.com/download/BrowserOS.dmg) · [Windows](https://files.browseros.com/download/BrowserOS_installer.exe) · [Linux (AppImage)](https://files.browseros.com/download/BrowserOS.AppImage) · [Linux (Debian)](https://cdn.browseros.com/download/BrowserOS.deb)
-2. **Import your Chrome data** (optional) — bookmarks, passwords, extensions all carry over
-3. **Connect your AI provider** — Claude, OpenAI, Gemini, ChatGPT Pro via OAuth, or local models via Ollama/LM Studio
+1. Download and install BrowserOS:
+   - [macOS](https://files.browseros.com/download/BrowserOS.dmg)
+   - [Windows](https://files.browseros.com/download/BrowserOS_installer.exe)
+   - [Linux (AppImage)](https://files.browseros.com/download/BrowserOS.AppImage)
+   - [Linux (Debian)](https://cdn.browseros.com/download/BrowserOS.deb)
 
-## Features
+2. Import your Chrome data (optional)
 
-| Feature | Description | Docs |
-|---------|-------------|------|
-| **AI Agent** | 53+ browser automation tools — navigate, click, type, extract data, all with natural language | [Guide](https://docs.browseros.com/getting-started) |
-| **MCP Server** | Control the browser from Claude Code, Gemini CLI, or any MCP client | [Setup](https://docs.browseros.com/features/use-with-claude-code) |
-| **Workflows** | Build repeatable browser automations with a visual graph builder | [Docs](https://docs.browseros.com/features/workflows) |
-| **Cowork** | Combine browser automation with local file operations — research the web, save reports to your folder | [Docs](https://docs.browseros.com/features/cowork) |
-| **Scheduled Tasks** | Run agents on autopilot — daily, hourly, or every few minutes | [Docs](https://docs.browseros.com/features/scheduled-tasks) |
-| **Memory** | Persistent memory across conversations — your assistant remembers context over time | [Docs](https://docs.browseros.com/features/memory) |
-| **SOUL.md** | Define your AI's personality and instructions in a single markdown file | [Docs](https://docs.browseros.com/features/soul-md) |
-| **LLM Hub** | Compare Claude, ChatGPT, and Gemini responses side-by-side on any page | [Docs](https://docs.browseros.com/features/llm-chat-hub) |
-| **40+ App Integrations** | Gmail, Slack, GitHub, Linear, Notion, Figma, Salesforce, and more via MCP | [Docs](https://docs.browseros.com/features/connect-apps) |
-| **Vertical Tabs** | Side-panel tab management — stay organized even with 100+ tabs open | [Docs](https://docs.browseros.com/features/vertical-tabs) |
-| **Ad Blocking** | uBlock Origin + Manifest V2 support — [10x more protection](https://docs.browseros.com/features/ad-blocking) than Chrome | [Docs](https://docs.browseros.com/features/ad-blocking) |
-| **Cloud Sync** | Sync browser config and agent history across devices | [Docs](https://docs.browseros.com/features/sync) |
-| **Skills** | Custom instruction sets that shape how your AI assistant behaves | [Docs](https://docs.browseros.com/features/skills) |
-| **Smart Nudges** | Contextual suggestions to connect apps and use features at the right moment | [Docs](https://docs.browseros.com/features/smart-nudges) |
+3. Connect your AI provider — use Claude, OpenAI, Gemini, or local models via Ollama and LMStudio.
+
+4. Start automating!
+
+## What makes BrowserOS special
+- 🏠 Feels like home — same Chrome interface, all your extensions just work
+- 🤖 AI agents that run on YOUR browser, not in the cloud
+- 🔒 Privacy first — bring your own keys or run local models with Ollama. Your browsing history stays on your machine
+- 🤝 [BrowserOS as MCP server](https://docs.browseros.com/features/use-with-claude-code) — control the browser from `claude-code`, `gemini-cli`, or any MCP client (31 tools)
+- 🔄 [Workflows](https://docs.browseros.com/features/workflows) — build repeatable browser automations with a visual graph builder
+- 📂 [Cowork](https://docs.browseros.com/features/cowork) — combine browser automation with local file operations. Research the web, save reports to your folder
+- ⏰ [Scheduled Tasks](https://docs.browseros.com/features/scheduled-tasks) — run the agent on autopilot, daily or every few minutes
+- 💬 [LLM Hub](https://docs.browseros.com/features/llm-chat-hub) — compare Claude, ChatGPT, and Gemini side-by-side on any page
+- 🛡️ Built-in ad blocker — [10x more protection than Chrome](https://docs.browseros.com/features/ad-blocking) with uBlock Origin + Manifest V2 support
+- 🚀 100% open source under AGPL-3.0
 
 ## Demos
 
-### BrowserOS agent in action
+### 🤖 BrowserOS agent in action
 [![BrowserOS agent in action](docs/videos/browserOS-agent-in-action.gif)](https://www.youtube.com/watch?v=SoSFev5R5dI)
 <br/><br/>
 
-### Install [BrowserOS as MCP](https://docs.browseros.com/features/use-with-claude-code) and control it from `claude-code`
+### 🎇 Install [BrowserOS as MCP](https://docs.browseros.com/features/use-with-claude-code) and control it from `claude-code`
 
 https://github.com/user-attachments/assets/c725d6df-1a0d-40eb-a125-ea009bf664dc
 
 <br/><br/>
 
-### Use BrowserOS to chat
+### 💬 Use BrowserOS to chat
 
 https://github.com/user-attachments/assets/726803c5-8e36-420e-8694-c63a2607beca
 
 <br/><br/>
 
-### Use BrowserOS to scrape data
+### ⚡ Use BrowserOS to scrape data
 
 https://github.com/user-attachments/assets/9f038216-bc24-4555-abf1-af2adcb7ebc0
 
 <br/><br/>
 
-## Install `browseros-cli`
+## Why We're Building BrowserOS
 
-Use `browseros-cli` to launch and control BrowserOS from the terminal or from AI coding agents like Claude Code.
+For the first time since Netscape pioneered the web in 1994, AI gives us the chance to completely reimagine the browser. We've seen tools like Cursor deliver 10x productivity gains for developers—yet everyday browsing remains frustratingly archaic.
 
-**macOS / Linux:**
+You're likely juggling 70+ tabs, battling your browser instead of having it assist you. Routine tasks, like ordering something from amazon or filling a form should be handled seamlessly by AI agents.
 
-```bash
-curl -fsSL https://cdn.browseros.com/cli/install.sh | bash
-```
+At BrowserOS, we're convinced that AI should empower you by automating tasks locally and securely—keeping your data private. We are building the best browser for this future!
 
-**Windows:**
+## How we compare
 
-```powershell
-irm https://cdn.browseros.com/cli/install.ps1 | iex
-```
+<details>
+<summary><b>vs Chrome</b></summary>
+<br>
+While we're grateful for Google open-sourcing Chromium, but Chrome hasn't evolved much in 10 years. No AI features, no automation, no MCP support.
+</details>
 
-After install, run `browseros-cli init` to connect the CLI to your running BrowserOS instance.
+<details>
+<summary><b>vs Brave</b></summary>
+<br>
+We love what Brave started, but they've spread themselves too thin with crypto, search, VPNs. We're laser-focused on AI-powered browsing.
+</details>
 
-## LLM Providers
+<details>
+<summary><b>vs Arc/Dia</b></summary>
+<br>
+Many loved Arc, but it was closed source. When they abandoned users, there was no recourse. We're 100% open source - fork it anytime!
+</details>
 
-BrowserOS works with any LLM. Bring your own keys, use OAuth, or run models locally.
+<details>
+<summary><b>vs Perplexity Comet</b></summary>
+<br>
+They're a search/ad company. Your browser history becomes their product. We keep everything local.
+</details>
 
-| Provider | Type | Auth |
-|----------|------|------|
-| Kimi K2.5 | Cloud (default) | Built-in |
-| ChatGPT Pro/Plus | Cloud | [OAuth](https://docs.browseros.com/features/chatgpt) |
-| GitHub Copilot | Cloud | [OAuth](https://docs.browseros.com/features/github-copilot) |
-| Qwen Code | Cloud | [OAuth](https://docs.browseros.com/features/qwen-code) |
-| Claude (Anthropic) | Cloud | API key |
-| GPT-4o / o3 (OpenAI) | Cloud | API key |
-| Gemini (Google) | Cloud | API key |
-| Azure OpenAI | Cloud | API key |
-| AWS Bedrock | Cloud | IAM credentials |
-| OpenRouter | Cloud | API key |
-| Ollama | Local | [Setup](https://docs.browseros.com/features/ollama) |
-| LM Studio | Local | [Setup](https://docs.browseros.com/features/lm-studio) |
-
-## How We Compare
-
-| | BrowserOS | Chrome | Brave | Dia | Comet | Atlas |
-|---|:---:|:---:|:---:|:---:|:---:|:---:|
-| Open Source | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ |
-| AI Agent | ✅ | ❌ | ❌ | ❌ | ✅ | ✅ |
-| MCP Server | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
-| Visual Workflows | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
-| Cowork (files + browser) | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
-| Scheduled Tasks | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ |
-| Bring Your Own Keys | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ |
-| Local Models (Ollama) | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ |
-| Local-first Privacy | ✅ | ❌ | ✅ | ❌ | ❌ | ❌ |
-| Ad Blocking (MV2) | ✅ | ❌ | ✅ | ❌ | ✅ | ❌ |
-
-**Detailed comparisons:**
-- [BrowserOS vs Chrome DevTools MCP](https://docs.browseros.com/comparisons/chrome-devtools-mcp) — developer-focused comparison for browser automation
-- [BrowserOS vs Claude Cowork](https://docs.browseros.com/comparisons/claude-cowork) — getting real work done with AI
-- [BrowserOS vs OpenClaw](https://docs.browseros.com/comparisons/openclaw) — everyday AI assistance
-
-## Architecture
-
-BrowserOS is a monorepo with two main subsystems: the **browser** (Chromium fork) and the **agent platform** (TypeScript/Go).
-
-```
-BrowserOS/
-├── packages/browseros/              # Chromium fork + build system (Python)
-│   ├── chromium_patches/            # Patches applied to Chromium source
-│   ├── build/                       # Build CLI and modules
-│   └── resources/                   # Icons, entitlements, signing
-│
-├── packages/browseros-agent/        # Agent platform (TypeScript/Go)
-│   ├── apps/
-│   │   ├── server/                  # MCP server + AI agent loop (Bun)
-│   │   ├── agent/                   # Browser extension UI (WXT + React)
-│   │   ├── cli/                     # CLI tool (Go)
-│   │   ├── eval/                    # Benchmark framework
-│   │   └── controller-ext/          # Chrome API bridge extension
-│   │
-│   └── packages/
-│       ├── agent-sdk/               # Node.js SDK (npm: @browseros-ai/agent-sdk)
-│       ├── cdp-protocol/            # CDP type bindings
-│       └── shared/                  # Shared constants
-```
-
-| Package | What it does |
-|---------|-------------|
-| [`packages/browseros`](packages/browseros/) | Chromium fork — patches, build system, signing |
-| [`apps/server`](packages/browseros-agent/apps/server/) | Bun server exposing 53+ MCP tools and running the AI agent loop |
-| [`apps/agent`](packages/browseros-agent/apps/agent/) | Browser extension — new tab, side panel chat, onboarding, settings |
-| [`apps/cli`](packages/browseros-agent/apps/cli/) | Go CLI — control BrowserOS from the terminal or AI coding agents |
-| [`apps/eval`](packages/browseros-agent/apps/eval/) | Benchmark framework — WebVoyager, Mind2Web evaluation |
-| [`agent-sdk`](packages/browseros-agent/packages/agent-sdk/) | Node.js SDK for browser automation with natural language |
-| [`cdp-protocol`](packages/browseros-agent/packages/cdp-protocol/) | Type-safe Chrome DevTools Protocol bindings |
+<details>
+<summary><b>vs ChatGPT Atlas</b></summary>
+<br>
+Your browsing data could be used for ads or to train their models. We keep your history and agent interactions strictly local.
+</details>
 
 ## Contributing
 
-We'd love your help making BrowserOS better! See our [Contributing Guide](CONTRIBUTING.md) for details.
+We'd love your help making BrowserOS better!
 
-- [Report bugs](https://github.com/browseros-ai/BrowserOS/issues)
-- [Suggest features](https://github.com/browseros-ai/BrowserOS/issues/99)
-- [Join Discord](https://discord.gg/YKwjt5vuKr) · [Join Slack](https://dub.sh/browserOS-slack)
-- [Follow on Twitter](https://x.com/browserOS_ai)
-
-**Agent development** (TypeScript/Go) — see the [agent monorepo README](packages/browseros-agent/README.md) for setup instructions.
-
-**Browser development** (C++/Python) — requires ~100GB disk space. See [`packages/browseros`](packages/browseros/) for build instructions.
-
-## Credits
-
-- [ungoogled-chromium](https://github.com/ungoogled-software/ungoogled-chromium) — BrowserOS uses some patches for enhanced privacy. Thanks to everyone behind this project!
-- [The Chromium Project](https://www.chromium.org/) — at the core of BrowserOS, making it possible to exist in the first place.
+- 🐛 [Report bugs](https://github.com/browseros-ai/BrowserOS/issues)
+- 💡 [Suggest features](https://github.com/browseros-ai/BrowserOS/issues/99)
+- 💬 [Join Discord](https://discord.gg/YKwjt5vuKr)
+- 🐦 [Follow on Twitter](https://x.com/browserOS_ai)
 
 ## License
 
 BrowserOS is open source under the [AGPL-3.0 license](LICENSE).
 
-Copyright &copy; 2026 Felafax, Inc.
+## Credits
+
+- [ungoogled-chromium](https://github.com/ungoogled-software/ungoogled-chromium) - BrowserOS uses some patches for enhanced privacy. Thanks to everyone behind this project!
+- [The Chromium Project](https://www.chromium.org/) - At the core of BrowserOS, making it possible to exist in the first place.
+
+## Citation
+
+If you use BrowserOS in your research or project, please cite:
+
+```bibtex
+@software{browseros2025,
+  author = {Sonti, Nithin and Sonti, Nikhil and {BrowserOS-team}},
+  title = {BrowserOS: The open-source Agentic browser},
+  url = {https://github.com/browseros-ai/BrowserOS},
+  year = {2025},
+  publisher = {GitHub},
+  license = {AGPL-3.0},
+}
+```
+
+Copyright &copy; 2025 Felafax, Inc.
 
 ## Stargazers
-
 Thank you to all our supporters!
 
 [![Star History Chart](https://api.star-history.com/svg?repos=browseros-ai/BrowserOS&type=Date)](https://www.star-history.com/#browseros-ai/BrowserOS&Date)
 
+## 
 <p align="center">
 Built with ❤️ from San Francisco
 </p>
+
+

packages/browseros-agent/.gitignore

@@ -195,4 +195,3 @@ test-results/
 .agent/
 .llm/
 .grove/
-docs/plans/2026-03-24-models-dev-integration.md

packages/browseros-agent/CLAUDE.md

@@ -32,7 +32,7 @@ Use **kebab-case** for all file and folder names:
 | Multi-word files | kebab-case | `gemini-agent.ts`, `mcp-context.ts` |
 | Single-word files | lowercase | `types.ts`, `browser.ts`, `index.ts` |
 | Test files | `.test.ts` suffix | `mcp-context.test.ts` |
-| Folders | kebab-case | `rate-limiter/`, `browser-tools/` |
+| Folders | kebab-case | `controller-server/`, `rate-limiter/` |
 
 Classes remain PascalCase in code, but live in kebab-case files:
 ```typescript
@@ -81,9 +81,6 @@ bun run dev:server               # Build server for development
 bun run dev:ext                  # Build extension for development
 bun run dist:server              # Build server for production (all targets)
 bun run dist:ext                 # Build extension for production
-
-# Refresh models.dev data
-bun run generate:models          # Fetches latest from models.dev/api.json
 ```
 
 ## Architecture
@@ -97,16 +94,21 @@ The main MCP server that exposes browser automation tools via HTTP/SSE.
 
 **Key components:**
 - `src/tools/` - MCP tool definitions, split into:
-  - `cdp-based/` - Tools using Chrome DevTools Protocol (navigation, DOM interaction, network, console, emulation, input, etc.)
+  - `cdp-based/` - Tools using Chrome DevTools Protocol (network, console, emulation, input, etc.)
+  - `controller-based/` - Tools using the browser extension (navigation, clicks, screenshots, tabs, history, bookmarks)
+- `src/controller-server/` - WebSocket server that bridges to the browser extension
+  - `ControllerBridge` handles WebSocket connections with extension clients
+  - `ControllerContext` wraps the bridge for tool handlers
 - `src/common/` - Shared utilities (McpContext, PageCollector, browser connection, identity, db)
 - `src/agent/` - AI agent functionality (Gemini adapter, rate limiting, session management)
 - `src/http/` - Hono HTTP server with MCP, health, and provider routes
 
 **Tool types:**
 - CDP tools require a direct CDP connection (`--cdp-port`)
+- Controller tools work via the browser extension over WebSocket
 
 ### Shared (`packages/shared`)
-Shared constants, types, and configuration used across packages. Avoids magic numbers.
+Shared constants, types, and configuration used by both server and extension. Avoids magic numbers.
 
 **Structure:**
 - `src/constants/` - Configuration values (ports, timeouts, limits, urls, paths)
@@ -114,12 +116,22 @@ Shared constants, types, and configuration used across packages. Avoids magic nu
 
 **Exports:** `@browseros/shared/constants/*`, `@browseros/shared/types/*`
 
+### Controller Extension (`apps/controller-ext`)
+Chrome extension that receives commands from the server via WebSocket.
+
+**Entry point:** `src/background/index.ts` → `BrowserOSController`
+
+**Structure:**
+- `src/actions/` - Action handlers organized by domain (browser/, tab/, bookmark/, history/)
+- `src/adapters/` - Chrome API adapters (TabAdapter, BookmarkAdapter, HistoryAdapter)
+- `src/websocket/` - WebSocket client that connects to the server
+
 ### Communication Flow
 
 ```
 AI Agent/MCP Client → HTTP Server (Hono) → Tool Handler
                                               ↓
-                                   CDP → BrowserOS / Chrome APIs
+                        CDP (direct) ←── or ──→ WebSocket → Extension → Chrome APIs
 ```
 
 ## Creating Packages

packages/browseros-agent/README.md

@@ -1,6 +1,8 @@
 # BrowserOS Agent
 
-The agent platform powering [BrowserOS](https://github.com/browseros-ai/BrowserOS) — contains the MCP server, agent UI, CLI, evaluation framework, and SDK.
+Monorepo for the BrowserOS-agent -- contains 3 packages: agent-UI, server (which contains the agent loop) and controller-extension (which is used by the tools within the agent loop).
+
+> **⚠️ NOTE:** This is only a submodule, the main project is at -- https://github.com/browseros-ai/BrowserOS
 
 ## Monorepo Structure
 
@@ -8,29 +10,24 @@ The agent platform powering [BrowserOS](https://github.com/browseros-ai/BrowserO
 apps/
   server/          # Bun server - MCP endpoints + agent loop
   agent/           # Agent UI (Chrome extension)
-  cli/             # Go CLI for controlling BrowserOS from the terminal
-  eval/            # Evaluation framework for benchmarking agents
+  controller-ext/  # BrowserOS Controller (Chrome extension for chrome.* APIs)
 
 packages/
-  agent-sdk/       # Node.js SDK (@browseros-ai/agent-sdk)
-  cdp-protocol/    # Type-safe Chrome DevTools Protocol bindings
   shared/          # Shared constants (ports, timeouts, limits)
 ```
 
 | Package | Description |
 |---------|-------------|
 | `apps/server` | Bun server exposing MCP tools and running the agent loop |
-| `apps/agent` | Agent UI — Chrome extension for the chat interface |
-| `apps/cli` | Go CLI — control BrowserOS from the terminal or AI coding agents |
-| `apps/eval` | Benchmark framework — WebVoyager, Mind2Web evaluation |
-| `packages/agent-sdk` | Node.js SDK for browser automation with natural language |
-| `packages/cdp-protocol` | Auto-generated CDP type bindings used by the server |
+| `apps/agent` | Agent UI - Chrome extension for the chat interface |
+| `apps/controller-ext` | BrowserOS Controller - Chrome extension that bridges `chrome.*` APIs (tabs, bookmarks, history) to the server via WebSocket |
 | `packages/shared` | Shared constants used across packages |
 
 ## Architecture
 
 - `apps/server`: Bun server which contains the agent loop and tools.
 - `apps/agent`: Agent UI (Chrome extension).
+- `apps/controller-ext`: BrowserOS Controller - a Chrome extension that bridges `chrome.*` APIs to the server. Controller tools within the server communicate with this extension via WebSocket.
 
 ```
 ┌──────────────────────────────────────────────────────────────────────────┐
@@ -48,19 +45,19 @@ packages/
 │   /health ─── Health check                                               │
 │                                                                          │
 │   Tools:                                                                 │
-│   └── CDP-backed browser tools (tabs, navigation, input, screenshots,   │
-│       bookmarks, history, console, DOM, tab groups, windows, ...)       │
+│   ├── CDP Tools (console, network, input, screenshot, ...)              │
+│   └── Controller Tools (tabs, navigation, clicks, bookmarks, history)   │
 └──────────────────────────────────────────────────────────────────────────┘
-                                    │
-                                    │ CDP (client)
-                                    ▼
-                         ┌─────────────────────┐
-                         │   Chromium CDP      │
-                         │  (cdpPort: 9000)    │
-                         │                     │
-                         │ Server connects     │
-                         │ TO this as client   │
-                         └─────────────────────┘
+          │                                         │
+          │ CDP (client)                            │ WebSocket (server)
+          ▼                                         ▼
+┌─────────────────────┐              ┌─────────────────────────────────────┐
+│   Chromium CDP      │              │   BrowserOS Controller Extension    │
+│  (cdpPort: 9000)    │              │     (extensionPort: 9300)           │
+│                     │              │                                     │
+│ Server connects     │              │ Bridges chrome.tabs, chrome.history │
+│ TO this as client   │              │ chrome.bookmarks to the server      │
+└─────────────────────┘              └─────────────────────────────────────┘
 ```
 
 ### Ports
@@ -69,7 +66,7 @@ packages/
 |------|--------------|---------|
 | 9100 | `BROWSEROS_SERVER_PORT` | HTTP server - MCP endpoints, agent chat, health |
 | 9000 | `BROWSEROS_CDP_PORT` | Chromium CDP server (BrowserOS Server connects as client) |
-| 9300 | `BROWSEROS_EXTENSION_PORT` | Legacy BrowserOS launch arg kept for compatibility; not used by the server |
+| 9300 | `BROWSEROS_EXTENSION_PORT` | WebSocket server for controller extension |
 
 ## Development
 
@@ -93,8 +90,9 @@ process-compose up
 
 The `process-compose up` command runs the following in order:
 1. `bun install` — installs dependencies
-2. `bun --cwd apps/agent codegen` — generates agent code
-3. `bun --cwd apps/server start` and `bun --cwd apps/agent dev` — starts server and agent in parallel
+2. `bun --cwd apps/controller-ext build` — builds the controller extension
+3. `bun --cwd apps/agent codegen` — generates agent code
+4. `bun --cwd apps/server start` and `bun --cwd apps/agent dev` — starts server and agent in parallel
 
 ### Environment Variables
 
@@ -110,7 +108,7 @@ Runtime uses `.env.development`, while production artifact builds use `.env.prod
 |----------|---------|-------------|
 | `BROWSEROS_SERVER_PORT` | 9100 | HTTP server port (MCP, chat, health) |
 | `BROWSEROS_CDP_PORT` | 9000 | Chromium CDP port (server connects as client) |
-| `BROWSEROS_EXTENSION_PORT` | 9300 | Legacy BrowserOS launch arg kept for compatibility |
+| `BROWSEROS_EXTENSION_PORT` | 9300 | WebSocket port for controller extension |
 | `BROWSEROS_CONFIG_URL` | - | Remote config endpoint for rate limits |
 | `BROWSEROS_INSTALL_ID` | - | Unique installation identifier (analytics) |
 | `BROWSEROS_CLIENT_ID` | - | Client identifier (analytics) |
@@ -142,7 +140,7 @@ Copy from `apps/server/.env.production.example` before running `build:server`.
 |----------|---------|-------------|
 | `BROWSEROS_SERVER_PORT` | 9100 | Passed to BrowserOS via CLI args |
 | `BROWSEROS_CDP_PORT` | 9000 | Passed to BrowserOS via CLI args |
-| `BROWSEROS_EXTENSION_PORT` | 9300 | Legacy BrowserOS CLI arg still passed for compatibility |
+| `BROWSEROS_EXTENSION_PORT` | 9300 | Passed to BrowserOS via CLI args |
 | `VITE_BROWSEROS_SERVER_PORT` | 9100 | Agent UI connects to server (must match `BROWSEROS_SERVER_PORT`) |
 | `BROWSEROS_BINARY` | - | Path to BrowserOS binary |
 | `USE_BROWSEROS_BINARY` | true | Use BrowserOS instead of default Chrome |
@@ -159,13 +157,15 @@ bun run start:server          # Start the server
 bun run start:agent           # Start agent extension (dev mode)
 
 # Build
-bun run build                 # Build server and agent
+bun run build                 # Build server, agent, and controller extension
 bun run build:server          # Build production server resource artifacts and upload zips to R2
 bun run build:agent           # Build agent extension
+bun run build:ext             # Build controller extension
 
 # Test
 bun run test                  # Run standard tests
 bun run test:cdp              # Run CDP-based tests
+bun run test:controller       # Run controller-based tests
 bun run test:integration      # Run integration tests
 
 # Quality

packages/browseros-agent/apps/agent/CHANGELOG.md

@@ -1,19 +0,0 @@
-# BrowserOS Agent Extension
-
-## v0.0.98 (2026-03-27)
-
-## What's Changed
-
-- chore: update agent version (#608)
-- chore: fix version number for extension (#606)
-- fix: improve chat history freshness and reduce query payload (#598)
-- feat: isolate new-tab agent navigation from origin tab (#593)
-- docs: overhaul READMEs across all major packages (#594)
-- fix(ui): resolve MCP promo banner dismiss button overlapping with text (#581)
-- docs: update agent extension changelog for v0.0.52 (#573)
-
-
-## v0.0.52 (2026-03-26)
-
-Initial release
-

packages/browseros-agent/apps/agent/README.md

@@ -1,24 +1,16 @@
-# BrowserOS Agent Extension
+# BrowserOS Agent Chrome Extension
 
-[![License: AGPL v3](https://img.shields.io/badge/License-AGPL%20v3-blue.svg)](../../../../LICENSE)
-
-The built-in browser extension that powers BrowserOS's AI interface — new tab with unified search, side panel chat, onboarding, and settings. Built with [WXT](https://wxt.dev) and React.
-
-> For user-facing feature documentation, see [docs.browseros.com](https://docs.browseros.com).
+The official Chrome extension for BrowserOS Agent, providing the UI layer for interacting with BrowserOS Core and Controllers. This extension enables intelligent browser automation, AI-powered search, and seamless integration with multiple LLM providers.
 
 ## Features
 
 - **AI-Powered New Tab**: Custom new tab page with unified search across Google and AI assistants
-- **Side Panel Chat**: Full-featured chat interface for interacting with BrowserOS
+- **Side Panel Chat**: Full-featured chat interface for interacting with BrowserOS Core
 - **Multi-Provider Support**: Connect to various LLM providers (OpenAI, Anthropic, Azure, Bedrock, and more)
 - **MCP Integration**: Model Context Protocol support for extending AI capabilities
 - **Visual Feedback**: Animated glow effect on tabs during AI agent operations
 - **Privacy-First**: Local data handling with configurable provider settings
 
-## How It Connects
-
-The extension communicates with the [BrowserOS Server](../../apps/server/) running locally. The server handles the AI agent loop, MCP tools, and CDP connections — the extension provides the UI layer.
-
 ## Project Structure
 
 ```
@@ -88,20 +80,47 @@ Settings dashboard with multiple sections:
 
 Content script that creates a visual indicator (pulsing orange glow) around the browser viewport when an AI agent is actively working on a tab.
 
+## How Tools Are Used
+
+### Bun
+
+Bun is the exclusive runtime and package manager:
+- All scripts use `bun run <script>` instead of npm
+- Package installation via `bun install`
+- Environment files automatically loaded (no dotenv needed)
+- Enforced via `engines` field in `package.json`
+
+```bash
+bun install          # Install dependencies
+bun run dev          # Development mode
+bun run build        # Production build
+bun run lint         # Run Biome linting
+```
+
+### Biome
+
+Unified linter and formatter configured in `biome.json`:
+- **Formatting**: 2-space indentation, single quotes, no semicolons
+- **Linting**: Recommended rules plus custom rules for unused imports/variables
+- **CSS Support**: Tailwind directives parsing enabled
+- **Import Organization**: Automatic import sorting via assist actions
+
+```bash
+bun run lint         # Check for issues
+bun run lint:fix     # Auto-fix issues
+```
+
 ## Development
 
 ### Prerequisites
 
 - [Bun](https://bun.sh) installed
 - Chrome or Chromium-based browser
-- BrowserOS Server running locally (for full functionality)
+- BrowserOS Core running locally (for full functionality)
 
 ### Setup
 
 ```bash
-# Copy environment file
-cp .env.example .env.development
-
 # Install dependencies
 bun install
 
@@ -134,30 +153,12 @@ SENTRY_AUTH_TOKEN=your-token
 
 ### GraphQL Schema
 
-Codegen requires a GraphQL schema. By default it uses the bundled `schema/schema.graphql`, so no extra setup is needed. If you have access to the original API source, you can set the following environment variable:
+Codegen requires a GraphQL schema. By default it uses the bundled `schema/schema.graphql`, so no extra setup is needed. If you have access to the original API source, you can set the following environment variable
 
 ```env
 GRAPHQL_SCHEMA_PATH=/path/to/api-repo/.../schema.graphql
 ```
 
-## Development Tooling
-
-### Bun
-
-Bun is the exclusive runtime and package manager:
-- All scripts use `bun run <script>` instead of npm
-- Package installation via `bun install`
-- Environment files automatically loaded (no dotenv needed)
-- Enforced via `engines` field in `package.json`
-
-### Biome
-
-Unified linter and formatter configured in `biome.json`:
-- **Formatting**: 2-space indentation, single quotes, no semicolons
-- **Linting**: Recommended rules plus custom rules for unused imports/variables
-- **CSS Support**: Tailwind directives parsing enabled
-- **Import Organization**: Automatic import sorting via assist actions
-
 ## Scripts
 
 | Script | Description |
@@ -168,5 +169,4 @@ Unified linter and formatter configured in `biome.json`:
 | `bun run lint` | Run Biome linter |
 | `bun run lint:fix` | Auto-fix linting issues |
 | `bun run typecheck` | Run TypeScript type checking |
-| `bun run codegen` | Generate GraphQL types |
 | `bun run clean:cache` | Clear build caches |

packages/browseros-agent/apps/agent/components/ai-elements/run-result-dialog.tsx

@@ -66,7 +66,7 @@ export const RunResultDialog: FC<RunResultDialogProps> = ({
 
   return (
     <Dialog open={!!run} onOpenChange={onOpenChange}>
-      <DialogContent className="sm:w-[70vw] sm:max-w-4xl">
+      <DialogContent className="sm:max-w-2xl">
         <DialogHeader>
           <DialogTitle className="flex items-center gap-2">
             {run.status === 'completed' ? (
@@ -94,7 +94,7 @@ export const RunResultDialog: FC<RunResultDialogProps> = ({
               <p className="text-destructive text-sm">{run.result}</p>
             </div>
           ) : run.result ? (
-            <div className="prose prose-sm dark:prose-invert [&_[data-streamdown='code-block']]:!w-full [&_[data-streamdown='table-wrapper']]:!w-full max-w-none break-words rounded-lg border border-border bg-muted/50 p-4 [&_[data-streamdown='table-wrapper']]:overflow-x-auto">
+            <div className="prose prose-sm dark:prose-invert [&_[data-streamdown='code-block']]:!w-full [&_[data-streamdown='table-wrapper']]:!w-full max-w-none break-words rounded-lg border border-border bg-muted/50 p-4">
               <MessageResponse>{run.result}</MessageResponse>
             </div>
           ) : (

packages/browseros-agent/apps/agent/components/credits/CreditBadge.tsx

@@ -14,7 +14,7 @@ export const CreditBadge: FC<CreditBadgeProps> = ({ credits, onClick }) => {
       type="button"
       onClick={onClick}
       className={cn(
-        'inline-flex cursor-pointer items-center gap-1 rounded-md px-1.5 py-0.5 font-medium text-xs transition-colors hover:bg-muted/50',
+        'inline-flex items-center gap-1 rounded-md px-1.5 py-0.5 font-medium text-xs transition-colors hover:bg-muted/50',
         getCreditTextColor(credits),
       )}
       title={`${credits} credits remaining`}

packages/browseros-agent/apps/agent/entrypoints/app/ai-settings/McpPromoBanner.tsx

@@ -17,7 +17,7 @@ export const McpPromoBanner: FC = () => {
   }
 
   return (
-    <div className="flex items-center gap-4 rounded-xl border border-border bg-card p-4 shadow-sm transition-all hover:shadow-md">
+    <div className="relative flex items-center gap-4 rounded-xl border border-border bg-card p-4 shadow-sm transition-all hover:shadow-md">
       <div className="flex h-10 w-10 shrink-0 items-center justify-center rounded-lg bg-[var(--accent-orange)]/10">
         <Server className="h-5 w-5 text-[var(--accent-orange)]" />
       </div>
@@ -48,7 +48,7 @@ export const McpPromoBanner: FC = () => {
       <button
         type="button"
         onClick={() => setDismissed(true)}
-        className="shrink-0 rounded-sm p-1 text-muted-foreground opacity-50 transition-opacity hover:opacity-100"
+        className="absolute top-2 right-2 rounded-sm p-1 text-muted-foreground opacity-50 transition-opacity hover:opacity-100"
       >
         <X className="h-3.5 w-3.5" />
       </button>

packages/browseros-agent/apps/agent/entrypoints/app/ai-settings/NewProviderDialog.tsx

@@ -1,26 +1,10 @@
 import { zodResolver } from '@hookform/resolvers/zod'
-import Fuse from 'fuse.js'
-import {
-  Check,
-  CheckCircle2,
-  ChevronDown,
-  ExternalLink,
-  Loader2,
-  XCircle,
-} from 'lucide-react'
-import { type FC, useEffect, useMemo, useState } from 'react'
+import { CheckCircle2, ExternalLink, Loader2, XCircle } from 'lucide-react'
+import { type FC, useEffect, useState } from 'react'
 import { useForm } from 'react-hook-form'
 import { z } from 'zod/v3'
 import { Button } from '@/components/ui/button'
 import { Checkbox } from '@/components/ui/checkbox'
-import {
-  Command,
-  CommandEmpty,
-  CommandGroup,
-  CommandInput,
-  CommandItem,
-  CommandList,
-} from '@/components/ui/command'
 import {
   Dialog,
   DialogContent,
@@ -39,11 +23,6 @@ import {
   FormMessage,
 } from '@/components/ui/form'
 import { Input } from '@/components/ui/input'
-import {
-  Popover,
-  PopoverContent,
-  PopoverTrigger,
-} from '@/components/ui/popover'
 import {
   Select,
   SelectContent,
@@ -56,10 +35,8 @@ import { useAgentServerUrl } from '@/lib/browseros/useBrowserOSProviders'
 import { useCapabilities } from '@/lib/browseros/useCapabilities'
 import {
   AI_PROVIDER_ADDED_EVENT,
-  AI_PROVIDER_UPDATED_EVENT,
   KIMI_API_KEY_CONFIGURED_EVENT,
   KIMI_API_KEY_GUIDE_CLICKED_EVENT,
-  MODEL_SELECTED_EVENT,
 } from '@/lib/constants/analyticsEvents'
 import { useKimiLaunch } from '@/lib/feature-flags/useKimiLaunch'
 import {
@@ -70,8 +47,7 @@ import {
 import { type TestResult, testProvider } from '@/lib/llm-providers/testProvider'
 import type { LlmProviderConfig, ProviderType } from '@/lib/llm-providers/types'
 import { track } from '@/lib/metrics/track'
-import { cn } from '@/lib/utils'
-import { getModelContextLength, getModelsForProvider } from './models'
+import { getModelContextLength, getModelOptions } from './models'
 
 const providerTypeEnum = z.enum([
   'moonshot',
@@ -187,13 +163,6 @@ export const providerFormSchema = z
  */
 export type ProviderFormValues = z.infer<typeof providerFormSchema>
 
-function formatContextWindow(tokens: number): string {
-  if (tokens >= 1000000)
-    return `${(tokens / 1000000).toFixed(tokens % 1000000 === 0 ? 0 : 1)}M`
-  if (tokens >= 1000) return `${Math.round(tokens / 1000)}K`
-  return `${tokens}`
-}
-
 /**
  * Props for NewProviderDialog
  * @public
@@ -219,10 +188,9 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
   initialValues,
   onSave,
 }) => {
+  const [isCustomModel, setIsCustomModel] = useState(false)
   const [isTesting, setIsTesting] = useState(false)
   const [testResult, setTestResult] = useState<TestResult | null>(null)
-  const [modelPickerOpen, setModelPickerOpen] = useState(false)
-  const [modelSearch, setModelSearch] = useState('')
   const { supports } = useCapabilities()
   const { baseUrl: agentServerUrl } = useAgentServerUrl()
   const kimiLaunch = useKimiLaunch()
@@ -293,21 +261,8 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
     watchedSessionToken,
   ])
 
-  const modelInfoList = getModelsForProvider(watchedType as ProviderType)
-
-  const modelFuse = useMemo(
-    () =>
-      new Fuse(modelInfoList, {
-        keys: ['modelId'],
-        threshold: 0.4,
-        distance: 100,
-      }),
-    [modelInfoList],
-  )
-
-  const filteredModels = modelSearch
-    ? modelFuse.search(modelSearch).map((r) => r.item)
-    : modelInfoList
+  // Get model options for current provider type
+  const modelOptions = getModelOptions(watchedType as ProviderType)
 
   // Handle provider type change (user-initiated via Select)
   const handleTypeChange = (newType: ProviderType) => {
@@ -317,13 +272,14 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
       form.setValue('baseUrl', defaultUrl)
     }
     form.setValue('modelId', '')
+    setIsCustomModel(false)
   }
 
   // Auto-fill context window when model changes (only for new providers)
   useEffect(() => {
     if (initialValues?.id) return
 
-    if (watchedModelId) {
+    if (watchedModelId && watchedModelId !== 'custom') {
       const contextLength = getModelContextLength(
         watchedType as ProviderType,
         watchedModelId,
@@ -334,6 +290,17 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
     }
   }, [watchedModelId, watchedType, form, initialValues?.id])
 
+  // Handle model selection (including custom option)
+  const handleModelChange = (value: string) => {
+    if (value === 'custom') {
+      setIsCustomModel(true)
+      form.setValue('modelId', '')
+    } else {
+      setIsCustomModel(false)
+      form.setValue('modelId', value)
+    }
+  }
+
   // Reset form when initialValues change
   useEffect(() => {
     if (initialValues) {
@@ -358,6 +325,7 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
         reasoningEffort: initialValues.reasoningEffort || 'high',
         reasoningSummary: initialValues.reasoningSummary || 'auto',
       })
+      setIsCustomModel(false)
     }
   }, [initialValues, form])
 
@@ -384,6 +352,7 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
         reasoningEffort: 'high',
         reasoningSummary: 'auto',
       })
+      setIsCustomModel(false)
     }
     // Clear test result when dialog opens/closes
     setTestResult(null)
@@ -404,11 +373,6 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
         provider_type: values.type,
         model: values.modelId,
       })
-    } else {
-      track(AI_PROVIDER_UPDATED_EVENT, {
-        provider_type: values.type,
-        model: values.modelId,
-      })
     }
     if (values.type === 'moonshot') {
       track(KIMI_API_KEY_CONFIGURED_EVENT, {
@@ -847,110 +811,52 @@ export const NewProviderDialog: FC<NewProviderDialogProps> = ({
               control={form.control}
               name="modelId"
               render={({ field }) => (
-                <FormItem className="flex flex-col">
+                <FormItem>
                   <FormLabel>Model *</FormLabel>
-                  {modelInfoList.length === 0 ? (
-                    <FormControl>
-                      <Input
-                        placeholder={
-                          watchedType === 'azure'
-                            ? 'Enter your deployment name'
-                            : watchedType === 'bedrock'
-                              ? 'e.g., anthropic.claude-3-5-sonnet-20241022-v2:0'
-                              : 'Enter model ID'
-                        }
-                        {...field}
-                      />
-                    </FormControl>
-                  ) : (
-                    <Popover
-                      open={modelPickerOpen}
-                      onOpenChange={(isOpen) => {
-                        setModelPickerOpen(isOpen)
-                        if (!isOpen) setModelSearch('')
-                      }}
-                    >
-                      <PopoverTrigger asChild>
-                        <button
+                  {isCustomModel || modelOptions.length === 1 ? (
+                    <>
+                      <FormControl>
+                        <Input
+                          placeholder={
+                            watchedType === 'azure'
+                              ? 'Enter your deployment name'
+                              : watchedType === 'bedrock'
+                                ? 'e.g., anthropic.claude-3-5-sonnet-20241022-v2:0'
+                                : 'Enter custom model ID'
+                          }
+                          {...field}
+                        />
+                      </FormControl>
+                      {modelOptions.length > 1 && (
+                        <Button
                           type="button"
-                          className={cn(
-                            'flex h-9 w-full items-center justify-between rounded-md border border-input bg-transparent px-3 py-1 text-sm shadow-xs',
-                            field.value
-                              ? 'text-foreground'
-                              : 'text-muted-foreground',
-                          )}
+                          variant="link"
+                          size="sm"
+                          className="h-auto p-0 text-xs"
+                          onClick={() => setIsCustomModel(false)}
                         >
-                          <span className="truncate">
-                            {field.value || 'Select a model...'}
-                          </span>
-                          <ChevronDown className="ml-2 h-4 w-4 shrink-0 opacity-50" />
-                        </button>
-                      </PopoverTrigger>
-                      <PopoverContent
-                        className="w-[var(--radix-popover-trigger-width)] p-0"
-                        align="start"
-                      >
-                        <Command shouldFilter={false}>
-                          <CommandInput
-                            placeholder="Search models..."
-                            value={modelSearch}
-                            onValueChange={setModelSearch}
-                            onKeyDown={(e) => {
-                              if (
-                                e.key === 'Enter' &&
-                                modelSearch &&
-                                filteredModels.length === 0
-                              ) {
-                                e.preventDefault()
-                                form.setValue('modelId', modelSearch)
-                                track(MODEL_SELECTED_EVENT, {
-                                  provider_type: watchedType,
-                                  model_id: modelSearch,
-                                  is_custom_model: true,
-                                })
-                                setModelPickerOpen(false)
-                                setModelSearch('')
-                              }
-                            }}
-                          />
-                          <CommandList>
-                            <CommandEmpty>
-                              No models found. Press Enter to use &quot;
-                              {modelSearch}&quot;
-                            </CommandEmpty>
-                            <CommandGroup>
-                              {filteredModels.map((model) => (
-                                <CommandItem
-                                  key={model.modelId}
-                                  value={model.modelId}
-                                  onSelect={() => {
-                                    form.setValue('modelId', model.modelId)
-                                    track(MODEL_SELECTED_EVENT, {
-                                      provider_type: watchedType,
-                                      model_id: model.modelId,
-                                      context_window: model.contextLength,
-                                      is_custom_model: false,
-                                    })
-                                    setModelPickerOpen(false)
-                                    setModelSearch('')
-                                  }}
-                                >
-                                  <span className="flex-1 truncate">
-                                    {model.modelId}
-                                  </span>
-                                  <span className="ml-2 shrink-0 rounded-md bg-muted px-1.5 py-0.5 font-mono text-[10px] text-muted-foreground">
-                                    {formatContextWindow(model.contextLength)}
-                                  </span>
-                                  {field.value === model.modelId && (
-                                    <Check className="ml-2 h-4 w-4 shrink-0" />
-                                  )}
-                                </CommandItem>
-                              ))}
-                            </CommandGroup>
-                          </CommandList>
-                        </Command>
-                      </PopoverContent>
-                    </Popover>
+                          ← Back to model list
+                        </Button>
+                      )}
+                    </>
+                  ) : (
+                    <Select
+                      onValueChange={handleModelChange}
+                      value={field.value}
+                    >
+                      <FormControl>
+                        <SelectTrigger className="w-full">
+                          <SelectValue placeholder="Select a model" />
+                        </SelectTrigger>
+                      </FormControl>
+                      <SelectContent>
+                        {modelOptions.map((modelId) => (
+                          <SelectItem key={modelId} value={modelId}>
+                            {modelId === 'custom' ? '+ Custom model' : modelId}
+                          </SelectItem>
+                        ))}
+                      </SelectContent>
+                    </Select>
                   )}
                   <FormMessage />
                 </FormItem>

packages/browseros-agent/apps/agent/entrypoints/app/ai-settings/models.ts

@@ -1,21 +1,98 @@
-import {
-  getModelsDevModels,
-  type ModelsDevModel,
-} from '@/lib/llm-providers/models-dev'
 import type { ProviderType } from '@/lib/llm-providers/types'
 
+/**
+ * Model information with context length
+ */
 export interface ModelInfo {
   modelId: string
   contextLength: number
-  supportsImages?: boolean
-  supportsReasoning?: boolean
-  supportsToolCall?: boolean
 }
 
-const CUSTOM_PROVIDER_MODELS: Partial<Record<ProviderType, ModelInfo[]>> = {
-  browseros: [{ modelId: 'browseros-auto', contextLength: 200000 }],
+/**
+ * Models data organized by provider type (matches backend AIProvider enum)
+ */
+export interface ModelsData {
+  anthropic: ModelInfo[]
+  openai: ModelInfo[]
+  'openai-compatible': ModelInfo[]
+  google: ModelInfo[]
+  openrouter: ModelInfo[]
+  azure: ModelInfo[]
+  ollama: ModelInfo[]
+  lmstudio: ModelInfo[]
+  bedrock: ModelInfo[]
+  browseros: ModelInfo[]
+  moonshot: ModelInfo[]
+  'chatgpt-pro': ModelInfo[]
+  'github-copilot': ModelInfo[]
+  'qwen-code': ModelInfo[]
+}
+
+/**
+ * Available models per provider with context lengths
+ * Based on: https://github.com/browseros-ai/BrowserOS-agent/blob/main/src/options/data/models.ts
+ */
+export const MODELS_DATA: ModelsData = {
+  moonshot: [{ modelId: 'kimi-k2.5', contextLength: 200000 }],
+  anthropic: [
+    { modelId: 'claude-opus-4-5-20251101', contextLength: 200000 },
+    { modelId: 'claude-haiku-4-5-20251001', contextLength: 200000 },
+    { modelId: 'claude-sonnet-4-5-20250929', contextLength: 200000 },
+    { modelId: 'claude-sonnet-4-20250514', contextLength: 200000 },
+    { modelId: 'claude-opus-4-20250514', contextLength: 200000 },
+    { modelId: 'claude-3-7-sonnet-20250219', contextLength: 200000 },
+    { modelId: 'claude-3-5-haiku-20241022', contextLength: 200000 },
+  ],
+  openai: [
+    { modelId: 'gpt-5.2', contextLength: 200000 },
+    { modelId: 'gpt-5.2-pro', contextLength: 200000 },
+    { modelId: 'gpt-5', contextLength: 200000 },
+    { modelId: 'gpt-5-mini', contextLength: 200000 },
+    { modelId: 'gpt-5-nano', contextLength: 200000 },
+    { modelId: 'gpt-4.1', contextLength: 200000 },
+    { modelId: 'gpt-4.1-mini', contextLength: 200000 },
+    { modelId: 'o4-mini', contextLength: 200000 },
+    { modelId: 'o3-mini', contextLength: 200000 },
+    { modelId: 'gpt-4o', contextLength: 128000 },
+    { modelId: 'gpt-4o-mini', contextLength: 128000 },
+  ],
   'openai-compatible': [],
-  ollama: [],
+  google: [
+    { modelId: 'gemini-3-pro-preview', contextLength: 1048576 },
+    { modelId: 'gemini-3-flash-preview', contextLength: 1048576 },
+    { modelId: 'gemini-2.5-flash', contextLength: 1048576 },
+    { modelId: 'gemini-2.5-pro', contextLength: 1048576 },
+  ],
+  openrouter: [
+    { modelId: 'google/gemini-3-pro-preview', contextLength: 1048576 },
+    { modelId: 'google/gemini-3-flash-preview', contextLength: 1048576 },
+    { modelId: 'google/gemini-2.5-flash', contextLength: 1048576 },
+    { modelId: 'anthropic/claude-opus-4.5', contextLength: 200000 },
+    { modelId: 'anthropic/claude-haiku-4.5', contextLength: 200000 },
+    { modelId: 'anthropic/claude-sonnet-4.5', contextLength: 200000 },
+    { modelId: 'anthropic/claude-sonnet-4', contextLength: 200000 },
+    { modelId: 'anthropic/claude-3.7-sonnet', contextLength: 200000 },
+    { modelId: 'openai/gpt-4o', contextLength: 128000 },
+    { modelId: 'openai/gpt-oss-120b', contextLength: 128000 },
+    { modelId: 'openai/gpt-oss-20b', contextLength: 128000 },
+    { modelId: 'qwen/qwen3-14b', contextLength: 131072 },
+    { modelId: 'qwen/qwen3-8b', contextLength: 131072 },
+  ],
+  azure: [],
+  ollama: [
+    { modelId: 'qwen3:4b', contextLength: 262144 },
+    { modelId: 'qwen3:8b', contextLength: 40960 },
+    { modelId: 'qwen3:14b', contextLength: 40960 },
+    { modelId: 'gpt-oss:20b', contextLength: 128000 },
+    { modelId: 'gpt-oss:120b', contextLength: 128000 },
+  ],
+  lmstudio: [
+    { modelId: 'openai/gpt-oss-20b', contextLength: 128000 },
+    { modelId: 'openai/gpt-oss-120b', contextLength: 128000 },
+    { modelId: 'qwen/qwen3-vl-8b', contextLength: 131072 },
+  ],
+  bedrock: [],
+  browseros: [{ modelId: 'browseros-auto', contextLength: 200000 }],
   'chatgpt-pro': [
     { modelId: 'gpt-5.4', contextLength: 400000 },
     { modelId: 'gpt-5.3-codex', contextLength: 400000 },
@@ -26,6 +103,32 @@ const CUSTOM_PROVIDER_MODELS: Partial<Record<ProviderType, ModelInfo[]>> = {
     { modelId: 'gpt-5.1-codex-mini', contextLength: 400000 },
     { modelId: 'gpt-5.1', contextLength: 200000 },
   ],
+  'github-copilot': [
+    // Free tier (unlimited with Pro)
+    { modelId: 'gpt-5-mini', contextLength: 128000 },
+    { modelId: 'claude-haiku-4.5', contextLength: 128000 },
+    { modelId: 'gpt-4o', contextLength: 64000 },
+    { modelId: 'gpt-4.1', contextLength: 64000 },
+    // Premium models (Pro: 300/mo, Pro+: 1500/mo)
+    { modelId: 'claude-sonnet-4.6', contextLength: 128000 },
+    { modelId: 'claude-sonnet-4.5', contextLength: 128000 },
+    { modelId: 'claude-sonnet-4', contextLength: 128000 },
+    { modelId: 'claude-opus-4.6', contextLength: 128000 },
+    { modelId: 'claude-opus-4.5', contextLength: 128000 },
+    { modelId: 'gemini-2.5-pro', contextLength: 128000 },
+    { modelId: 'gemini-3-pro-preview', contextLength: 128000 },
+    { modelId: 'gemini-3-flash-preview', contextLength: 128000 },
+    { modelId: 'gemini-3.1-pro-preview', contextLength: 128000 },
+    { modelId: 'gpt-5.4', contextLength: 272000 },
+    { modelId: 'gpt-5.4-mini', contextLength: 128000 },
+    { modelId: 'gpt-5.3-codex', contextLength: 272000 },
+    { modelId: 'gpt-5.2-codex', contextLength: 272000 },
+    { modelId: 'gpt-5.2', contextLength: 128000 },
+    { modelId: 'gpt-5.1-codex', contextLength: 128000 },
+    { modelId: 'gpt-5.1-codex-max', contextLength: 128000 },
+    { modelId: 'gpt-5.1', contextLength: 128000 },
+    { modelId: 'grok-code-fast-1', contextLength: 128000 },
+  ],
   'qwen-code': [
     { modelId: 'coder-model', contextLength: 1000000 },
     { modelId: 'qwen3-coder-plus', contextLength: 1000000 },
@@ -34,23 +137,25 @@ const CUSTOM_PROVIDER_MODELS: Partial<Record<ProviderType, ModelInfo[]>> = {
   ],
 }
 
-function fromModelsDevModel(m: ModelsDevModel): ModelInfo {
-  return {
-    modelId: m.id,
-    contextLength: m.contextWindow,
-    supportsImages: m.supportsImages,
-    supportsReasoning: m.supportsReasoning,
-    supportsToolCall: m.supportsToolCall,
-  }
-}
-
+/**
+ * Get models for a specific provider type
+ */
 export function getModelsForProvider(providerType: ProviderType): ModelInfo[] {
-  const custom = CUSTOM_PROVIDER_MODELS[providerType]
-  if (custom !== undefined) return custom
-
-  return getModelsDevModels(providerType).map(fromModelsDevModel)
+  return MODELS_DATA[providerType] || []
 }
 
+/**
+ * Get model options for select dropdown (model IDs + custom option)
+ */
+export function getModelOptions(providerType: ProviderType): string[] {
+  const models = getModelsForProvider(providerType)
+  const modelIds = models.map((m) => m.modelId)
+  return modelIds.length > 0 ? [...modelIds, 'custom'] : ['custom']
+}
+
+/**
+ * Get context length for a specific model
+ */
 export function getModelContextLength(
   providerType: ProviderType,
   modelId: string,
@@ -59,3 +164,14 @@ export function getModelContextLength(
   const model = models.find((m) => m.modelId === modelId)
   return model?.contextLength
 }
+
+/**
+ * Check if model ID is a custom (user-entered) value
+ */
+export function isCustomModel(
+  providerType: ProviderType,
+  modelId: string,
+): boolean {
+  const models = getModelsForProvider(providerType)
+  return !models.some((m) => m.modelId === modelId)
+}

packages/browseros-agent/apps/agent/entrypoints/app/login/LogoutPage.tsx

@@ -1,5 +1,5 @@
 import { useQueryClient } from '@tanstack/react-query'
-import { clear } from 'idb-keyval'
+import localforage from 'localforage'
 import { Loader2 } from 'lucide-react'
 import type { FC } from 'react'
 import { useEffect } from 'react'
@@ -25,7 +25,7 @@ export const LogoutPage: FC = () => {
       await providersStorage.removeValue()
       await scheduledJobStorage.removeValue()
       queryClient.clear()
-      await clear()
+      await localforage.clear()
 
       resetIdentity()
       await signOut()

packages/browseros-agent/apps/agent/entrypoints/newtab/index/NewTabChat.tsx

@@ -169,15 +169,8 @@ export const NewTabChat: FC = () => {
             onDismissJtbdPopup={() => {}}
           />
         )}
-        {agentUrlError && (
-          <ChatError
-            error={agentUrlError}
-            providerType={selectedProvider?.type}
-          />
-        )}
-        {chatError && (
-          <ChatError error={chatError} providerType={selectedProvider?.type} />
-        )}
+        {agentUrlError && <ChatError error={agentUrlError} />}
+        {chatError && <ChatError error={chatError} />}
       </main>
 
       <div className="mx-auto w-full max-w-3xl flex-shrink-0 px-4 pb-2">

packages/browseros-agent/apps/agent/entrypoints/sidepanel/history/ChatHistory.tsx

@@ -32,7 +32,6 @@ const RemoteChatHistory: FC<{ userId: string }> = ({ userId }) => {
   const {
     data: graphqlData,
     isLoading: isLoadingConversations,
-    isFetching,
     hasNextPage,
     isFetchingNextPage,
     fetchNextPage,
@@ -113,7 +112,6 @@ const RemoteChatHistory: FC<{ userId: string }> = ({ userId }) => {
       hasNextPage={hasNextPage}
       isFetchingNextPage={isFetchingNextPage}
       onLoadMore={fetchNextPage}
-      isRefreshing={isFetching && !isLoadingConversations}
     />
   )
 }

packages/browseros-agent/apps/agent/entrypoints/sidepanel/history/components/ConversationList.tsx

@@ -12,7 +12,6 @@ interface ConversationListProps {
   hasNextPage?: boolean
   isFetchingNextPage?: boolean
   onLoadMore?: () => void
-  isRefreshing?: boolean
 }
 
 export const ConversationList: FC<ConversationListProps> = ({
@@ -22,7 +21,6 @@ export const ConversationList: FC<ConversationListProps> = ({
   hasNextPage,
   isFetchingNextPage,
   onLoadMore,
-  isRefreshing,
 }) => {
   const loadMoreRef = useRef<HTMLDivElement>(null)
 
@@ -59,12 +57,6 @@ export const ConversationList: FC<ConversationListProps> = ({
   return (
     <main className="mt-4 flex h-full flex-1 flex-col space-y-4 overflow-y-auto">
       <div className="w-full p-3">
-        {isRefreshing && (
-          <div className="flex items-center justify-center gap-2 pb-3 text-muted-foreground text-xs">
-            <Loader2 className="h-3 w-3 animate-spin" />
-            <span>Fetching latest conversations</span>
-          </div>
-        )}
         {!hasConversations ? (
           <div className="flex flex-col items-center justify-center py-12 text-center">
             <MessageSquare className="mb-3 h-10 w-10 text-muted-foreground/50" />

packages/browseros-agent/apps/agent/entrypoints/sidepanel/history/graphql/chatHistoryDocument.ts

@@ -11,7 +11,7 @@ export const GetConversationsForHistoryDocument = graphql(`
       nodes {
         rowId
         lastMessagedAt
-        conversationMessages(first: 2, orderBy: ORDER_INDEX_DESC) {
+        conversationMessages(last: 5, orderBy: ORDER_INDEX_ASC) {
           nodes {
             message
           }

packages/browseros-agent/apps/agent/entrypoints/sidepanel/index/Chat.tsx

@@ -224,12 +224,7 @@ export const Chat = () => {
             onDismissJtbdPopup={onDismissJtbdPopup}
           />
         )}
-        {agentUrlError && (
-          <ChatError
-            error={agentUrlError}
-            providerType={selectedProvider?.type}
-          />
-        )}
+        {agentUrlError && <ChatError error={agentUrlError} />}
         {chatError && (
           <ChatError error={chatError} providerType={selectedProvider?.type} />
         )}

packages/browseros-agent/apps/agent/entrypoints/sidepanel/index/ChatError.tsx

@@ -34,9 +34,11 @@ function parseErrorMessage(
 } {
   const isBrowserosProvider = providerType === 'browseros'
 
-  // All chat requests go through the local BrowserOS agent server, so any
-  // fetch failure is always a local connection issue.
-  if (message.includes('Failed to fetch') || message.includes('fetch failed')) {
+  // Detect MCP server connection failures (universal — affects all providers)
+  if (
+    (message.includes('Failed to fetch') || message.includes('fetch failed')) &&
+    message.includes('127.0.0.1')
+  ) {
     return {
       text: 'Unable to connect to BrowserOS agent. Follow below instructions.',
       url: 'https://docs.browseros.com/troubleshooting/connection-issues',

packages/browseros-agent/apps/agent/entrypoints/sidepanel/index/useChatSession.ts

@@ -76,6 +76,8 @@ export interface ChatSessionOptions {
   isIntegrationsSynced?: boolean
 }
 
+const NEWTAB_SYSTEM_PROMPT = `IMPORTANT: The user is chatting from the New Tab page. When performing browser actions, ALWAYS open content in a NEW TAB rather than navigating the current tab. The user's new tab page should remain accessible.`
+
 export const useChatSession = (options?: ChatSessionOptions) => {
   const {
     selectedLlmProviderRef,
@@ -342,8 +344,12 @@ export const useChatSession = (options?: ChatSessionOptions) => {
             reasoningEffort: provider?.reasoningEffort,
             reasoningSummary: provider?.reasoningSummary,
             browserContext,
-            origin: options?.origin ?? 'sidepanel',
-            userSystemPrompt: personalizationRef.current,
+            userSystemPrompt:
+              options?.origin === 'newtab'
+                ? [personalizationRef.current, NEWTAB_SYSTEM_PROMPT]
+                    .filter(Boolean)
+                    .join('\n\n')
+                : personalizationRef.current,
             userWorkingDir: workingDirRef.current,
             supportsImages: provider?.supportsImages,
             previousConversation,
@@ -561,11 +567,9 @@ export const useChatSession = (options?: ChatSessionOptions) => {
   }, [])
 
   const handleSelectProvider = (provider: Provider) => {
-    const fullProvider = llmProviders.find((p) => p.id === provider.id)
     track(PROVIDER_SELECTED_EVENT, {
       provider_id: provider.id,
       provider_type: provider.type,
-      model_id: fullProvider?.modelId,
     })
     setDefaultProvider(provider.id)
   }

packages/browseros-agent/apps/agent/lib/constants/analyticsEvents.ts

@@ -29,12 +29,6 @@ export const CONVERSATION_RESET_EVENT = 'ui.conversation.reset'
 /** @public */
 export const AI_PROVIDER_ADDED_EVENT = 'settings.ai_provider.added'
 
-/** @public */
-export const AI_PROVIDER_UPDATED_EVENT = 'settings.ai_provider.updated'
-
-/** @public */
-export const MODEL_SELECTED_EVENT = 'settings.model.selected'
-
 /** @public */
 export const CHATGPT_PRO_OAUTH_STARTED_EVENT =
   'settings.chatgpt_pro.oauth_started'

packages/browseros-agent/apps/agent/lib/graphql/QueryProvider.tsx

@@ -1,10 +1,7 @@
 import { createAsyncStoragePersister } from '@tanstack/query-async-storage-persister'
 import { QueryClient } from '@tanstack/react-query'
-import {
-  type AsyncStorage,
-  PersistQueryClientProvider,
-} from '@tanstack/react-query-persist-client'
-import { del, get, set } from 'idb-keyval'
+import { PersistQueryClientProvider } from '@tanstack/react-query-persist-client'
+import localforage from 'localforage'
 import type { FC, ReactNode } from 'react'
 
 const queryClient = new QueryClient({
@@ -15,14 +12,8 @@ const queryClient = new QueryClient({
   },
 })
 
-const idbStorage: AsyncStorage<string> = {
-  getItem: (key: string) => get<string>(key).then((v) => v ?? null),
-  setItem: (key: string, value: string) => set(key, value),
-  removeItem: (key: string) => del(key),
-}
-
 const asyncStoragePersister = createAsyncStoragePersister({
-  storage: idbStorage,
+  storage: localforage,
 })
 
 export const QueryProvider: FC<{ children: ReactNode }> = ({ children }) => {

packages/browseros-agent/apps/agent/lib/llm-providers/models-dev.ts

@@ -1,35 +0,0 @@
-import data from './models-dev-data.json'
-
-export interface ModelsDevModel {
-  id: string
-  name: string
-  contextWindow: number
-  maxOutput: number
-  supportsImages: boolean
-  supportsReasoning: boolean
-  supportsToolCall: boolean
-  inputCost?: number
-  outputCost?: number
-}
-
-export interface ModelsDevProvider {
-  name: string
-  api?: string
-  doc: string
-  models: ModelsDevModel[]
-}
-
-const modelsDevData: Record<string, ModelsDevProvider> = data as Record<
-  string,
-  ModelsDevProvider
->
-
-export function getModelsDevProvider(
-  providerId: string,
-): ModelsDevProvider | undefined {
-  return modelsDevData[providerId]
-}
-
-export function getModelsDevModels(providerId: string): ModelsDevModel[] {
-  return modelsDevData[providerId]?.models ?? []
-}

packages/browseros-agent/apps/agent/lib/llm-providers/providerTemplates.ts

@@ -1,4 +1,3 @@
-import { getModelsDevProvider } from './models-dev'
 import type { ProviderType } from './types'
 
 /**
@@ -16,30 +15,6 @@ export interface ProviderTemplate {
   apiKeyUrl?: string
 }
 
-function enrichTemplate(
-  providerId: ProviderType,
-  overrides: {
-    defaultModelId: string
-    defaultBaseUrl?: string
-    apiKeyUrl?: string
-    setupGuideUrl?: string
-  },
-): ProviderTemplate {
-  const provider = getModelsDevProvider(providerId)
-  const model = provider?.models.find((m) => m.id === overrides.defaultModelId)
-
-  return {
-    id: providerId,
-    name: provider?.name ?? providerId,
-    defaultBaseUrl: overrides.defaultBaseUrl ?? provider?.api ?? '',
-    defaultModelId: overrides.defaultModelId,
-    supportsImages: model?.supportsImages ?? true,
-    contextWindow: model?.contextWindow ?? 128000,
-    ...(overrides.apiKeyUrl && { apiKeyUrl: overrides.apiKeyUrl }),
-    ...(overrides.setupGuideUrl && { setupGuideUrl: overrides.setupGuideUrl }),
-  }
-}
-
 /**
  * Available provider templates for quick setup
  * @public
@@ -82,12 +57,17 @@ export const providerTemplates: ProviderTemplate[] = [
     apiKeyUrl: 'https://platform.moonshot.ai/console/api-keys',
     setupGuideUrl: 'https://platform.moonshot.ai/console/api-keys',
   },
-  enrichTemplate('openai', {
-    defaultModelId: 'gpt-5',
+  {
+    id: 'openai',
+    name: 'OpenAI',
+    defaultBaseUrl: 'https://api.openai.com/v1',
+    defaultModelId: 'gpt-4',
+    supportsImages: true,
+    contextWindow: 128000,
     apiKeyUrl: 'https://platform.openai.com/api-keys',
     setupGuideUrl:
       'https://docs.browseros.com/features/bring-your-own-llm#openai',
-  }),
+  },
   {
     id: 'openai-compatible',
     name: 'OpenAI Compatible',
@@ -96,18 +76,28 @@ export const providerTemplates: ProviderTemplate[] = [
     supportsImages: true,
     contextWindow: 128000,
   },
-  enrichTemplate('anthropic', {
-    defaultModelId: 'claude-sonnet-4-6',
+  {
+    id: 'anthropic',
+    name: 'Anthropic',
+    defaultBaseUrl: 'https://api.anthropic.com/v1',
+    defaultModelId: 'claude-3-5-sonnet-20241022',
+    supportsImages: true,
+    contextWindow: 200000,
     apiKeyUrl: 'https://console.anthropic.com/settings/keys',
     setupGuideUrl:
       'https://docs.browseros.com/features/bring-your-own-llm#claude',
-  }),
-  enrichTemplate('google', {
-    defaultModelId: 'gemini-2.5-flash',
+  },
+  {
+    id: 'google',
+    name: 'Gemini',
+    defaultBaseUrl: 'https://generativelanguage.googleapis.com/v1beta',
+    defaultModelId: 'gemini-1.5-pro',
+    supportsImages: true,
+    contextWindow: 1000000,
     apiKeyUrl: 'https://aistudio.google.com/app/apikey',
     setupGuideUrl:
       'https://docs.browseros.com/features/bring-your-own-llm#gemini',
-  }),
+  },
   {
     id: 'ollama',
     name: 'Ollama',
@@ -118,28 +108,47 @@ export const providerTemplates: ProviderTemplate[] = [
     setupGuideUrl:
       'https://docs.browseros.com/features/bring-your-own-llm#ollama',
   },
-  enrichTemplate('openrouter', {
-    defaultModelId: 'anthropic/claude-sonnet-4.5',
+  {
+    id: 'openrouter',
+    name: 'OpenRouter',
+    defaultBaseUrl: 'https://openrouter.ai/api/v1',
+    defaultModelId: 'openai/gpt-4-turbo',
+    supportsImages: true,
+    contextWindow: 128000,
     apiKeyUrl: 'https://openrouter.ai/keys',
     setupGuideUrl:
       'https://docs.browseros.com/features/bring-your-own-llm#openrouter',
-  }),
-  enrichTemplate('lmstudio', {
-    defaultModelId: 'openai/gpt-oss-20b',
+  },
+  {
+    id: 'lmstudio',
+    name: 'LM Studio',
     defaultBaseUrl: 'http://localhost:1234/v1',
+    defaultModelId: 'local-model',
+    supportsImages: false,
+    contextWindow: 32000,
     setupGuideUrl:
       'https://docs.browseros.com/features/bring-your-own-llm#lmstudio',
-  }),
-  enrichTemplate('azure', {
+  },
+  {
+    id: 'azure',
+    name: 'Azure',
+    defaultBaseUrl: '',
     defaultModelId: '',
+    supportsImages: true,
+    contextWindow: 128000,
     apiKeyUrl:
       'https://portal.azure.com/#view/Microsoft_Azure_ProjectOxford/CognitiveServicesHub/~/OpenAI',
-  }),
-  enrichTemplate('bedrock', {
-    defaultModelId: 'anthropic.claude-sonnet-4-6',
+  },
+  {
+    id: 'bedrock',
+    name: 'AWS Bedrock',
+    defaultBaseUrl: '',
+    defaultModelId: '',
+    supportsImages: true,
+    contextWindow: 200000,
     setupGuideUrl:
       'https://docs.aws.amazon.com/bedrock/latest/userguide/getting-started.html',
-  }),
+  },
 ]
 
 /**

packages/browseros-agent/apps/agent/package.json

@@ -2,7 +2,7 @@
   "name": "@browseros/agent",
   "description": "manifest.json description",
   "private": true,
-  "version": "0.0.98",
+  "version": "0.0.52",
   "type": "module",
   "scripts": {
     "dev": "test -d generated/graphql || bun run codegen; mkdir -p /tmp/browseros-dev; bun --env-file=.env.development wxt",
@@ -44,9 +44,9 @@
     "@radix-ui/react-use-controllable-state": "^1.2.2",
     "@sentry/react": "^10.31.0",
     "@sentry/vite-plugin": "^4.6.1",
-    "@tanstack/query-async-storage-persister": "^5.95.2",
-    "@tanstack/react-query": "^5.95.2",
-    "@tanstack/react-query-persist-client": "^5.95.2",
+    "@tanstack/query-async-storage-persister": "^5.90.21",
+    "@tanstack/react-query": "^5.90.19",
+    "@tanstack/react-query-persist-client": "^5.90.21",
     "@types/cytoscape": "^3.31.0",
     "@types/dompurify": "^3.2.0",
     "@webext-core/messaging": "^2.3.0",
@@ -67,11 +67,10 @@
     "embla-carousel-react": "^8.6.0",
     "es-toolkit": "^1.42.0",
     "eventsource-parser": "^3.0.6",
-    "fuse.js": "^7.1.0",
     "graphql": "^16.12.0",
     "hono": "^4.12.3",
-    "idb-keyval": "^6.2.2",
     "klavis": "^2.15.0",
+    "localforage": "^1.10.0",
     "lucide-react": "^0.562.0",
     "motion": "^12.23.24",
     "nanoid": "^5.1.6",

packages/browseros-agent/apps/agent/web-ext.config.ts

@@ -1,13 +1,19 @@
+import { dirname, join } from 'node:path'
+import { fileURLToPath } from 'node:url'
 import { defineWebExtConfig } from 'wxt'
 
 // biome-ignore lint/style/noProcessEnv: config file needs env access
 const env = process.env
 
+const MONOREPO_ROOT = join(dirname(fileURLToPath(import.meta.url)), '../..')
+const CONTROLLER_EXT_DIR = join(MONOREPO_ROOT, 'apps/controller-ext/dist')
+
 const chromiumArgs = [
   '--use-mock-keychain',
   '--show-component-extension-options',
   '--disable-browseros-server',
   '--disable-browseros-extensions',
+  `--load-extension=${CONTROLLER_EXT_DIR}`,
 ]
 
 if (env.BROWSEROS_CDP_PORT) {

packages/browseros-agent/apps/cli/.env.production.example

@@ -1,10 +0,0 @@
-# Production build env for CLI
-
-POSTHOG_API_KEY=
-
-# Upload env for CLI installer scripts
-R2_ACCOUNT_ID=
-R2_ACCESS_KEY_ID=
-R2_SECRET_ACCESS_KEY=
-R2_BUCKET=browseros
-R2_UPLOAD_PREFIX=cli

packages/browseros-agent/apps/cli/.goreleaser.yml

@@ -0,0 +1,47 @@
+version: 2
+
+project_name: browseros-cli
+
+builds:
+  - main: .
+    binary: browseros-cli
+    env:
+      - CGO_ENABLED=0
+    flags:
+      - -trimpath
+    ldflags:
+      - -s -w -X main.version={{ .Version }}
+    targets:
+      - darwin_amd64
+      - darwin_arm64
+      - linux_amd64
+      - linux_arm64
+      - windows_amd64
+      - windows_arm64
+
+archives:
+  - format: tar.gz
+    format_overrides:
+      - goos: windows
+        format: zip
+    name_template: "{{ .ProjectName }}_{{ .Version }}_{{ .Os }}_{{ .Arch }}"
+    files:
+      - "none*"
+
+checksum:
+  name_template: checksums.txt
+
+changelog:
+  sort: asc
+  filters:
+    exclude:
+      - "^docs:"
+      - "^test:"
+      - "^ci:"
+
+release:
+  github:
+    owner: browseros-ai
+    name: BrowserOS
+  prerelease: auto
+  name_template: "browseros-cli v{{ .Version }}"

packages/browseros-agent/apps/cli/CHANGELOG.md

@@ -1 +0,0 @@
-# BrowserOS CLI

packages/browseros-agent/apps/cli/Makefile

@@ -1,27 +1,17 @@
 BINARY := browseros-cli
 SOURCES := $(shell find . -name '*.go')
 VERSION ?= dev
-POSTHOG_API_KEY ?=
-DIST := dist
-LDFLAGS := -X main.version=$(VERSION) -X browseros-cli/analytics.posthogAPIKey=$(POSTHOG_API_KEY)
-HOST_OS := $(shell go env GOOS)
-HOST_ARCH := $(shell go env GOARCH)
-HOST_EXT := $(if $(filter windows,$(HOST_OS)),.exe,)
-HOST_BINARY = $(DIST)/$(BINARY)_$(HOST_OS)_$(HOST_ARCH)$(HOST_EXT)
 
 $(BINARY): $(SOURCES)
-	go build -ldflags "$(LDFLAGS)" -o $(BINARY) .
+	go build -ldflags "-X main.version=$(VERSION)" -o $(BINARY) .
 
-PLATFORMS := darwin/amd64 darwin/arm64 linux/amd64 linux/arm64 windows/amd64 windows/arm64
-
-.PHONY: install clean vet test release
+.PHONY: install clean vet test
 
 install:
-	go install -ldflags "$(LDFLAGS)" .
+	go install -ldflags "-X main.version=$(VERSION)" .
 
 clean:
 	rm -f $(BINARY)
-	rm -rf $(DIST)
 
 vet:
 	go vet ./...
@@ -29,41 +19,8 @@ vet:
 test:
 	go test -tags integration -v -timeout 120s ./...
 
+release-dry:
+	goreleaser release --snapshot --clean
+
 release:
-	@if [ "$(VERSION)" = "dev" ]; then echo "Error: VERSION required (e.g. make release VERSION=0.1.0)" >&2; exit 1; fi
-	@rm -rf $(DIST) && mkdir -p $(DIST)
-	@for pair in $(PLATFORMS); do \
-		OS=$${pair%/*}; \
-		ARCH=$${pair#*/}; \
-		EXT=""; \
-		if [ "$$OS" = "windows" ]; then EXT=".exe"; fi; \
-		echo "Building $$OS/$$ARCH..."; \
-		GOOS=$$OS GOARCH=$$ARCH CGO_ENABLED=0 go build -trimpath \
-			-ldflags "-s -w $(LDFLAGS)" \
-			-o "$(DIST)/$(BINARY)$$EXT" .; \
-		ARCHIVE="$(BINARY)_$(VERSION)_$${OS}_$${ARCH}"; \
-		if [ "$$OS" = "windows" ]; then \
-			(cd $(DIST) && zip "$${ARCHIVE}.zip" "$(BINARY)$$EXT"); \
-		else \
-			(cd $(DIST) && tar czf "$${ARCHIVE}.tar.gz" "$(BINARY)"); \
-		fi; \
-		mv "$(DIST)/$(BINARY)$$EXT" "$(DIST)/$(BINARY)_$${OS}_$${ARCH}$$EXT"; \
-	done
-	@ACTUAL_VERSION=$$($(HOST_BINARY) --version | awk '{print $$3}'); \
-		if [ "$$ACTUAL_VERSION" != "$(VERSION)" ]; then \
-			echo "Error: expected $(HOST_BINARY) to report version $(VERSION), got $$ACTUAL_VERSION" >&2; \
-			exit 1; \
-		fi
-	@cd $(DIST) && (command -v sha256sum >/dev/null 2>&1 && sha256sum *.tar.gz *.zip || shasum -a 256 *.tar.gz *.zip) > checksums.txt
-	@echo "=== Built artifacts ==="
-	@ls -lh $(DIST)
-
-.PHONY: npm-version npm-publish
-
-npm-version:
-	@if [ "$(VERSION)" = "dev" ]; then echo "Error: VERSION required" >&2; exit 1; fi
-	@node -e "const p=require('./npm/package.json');p.version='$(VERSION)';require('fs').writeFileSync('./npm/package.json',JSON.stringify(p,null,2)+'\n')"
-	@echo "npm/package.json version set to $(VERSION)"
-
-npm-publish: npm-version
-	cd npm && npm publish
+	goreleaser release --clean

packages/browseros-agent/apps/cli/README.md

@@ -1,68 +1,25 @@
 # browseros-cli
 
-[![License: AGPL v3](https://img.shields.io/badge/License-AGPL%20v3-blue.svg)](../../../../LICENSE)
+Command-line interface for controlling BrowserOS via MCP. Talks to the BrowserOS MCP server over JSON-RPC 2.0 / StreamableHTTP.
 
-Command-line interface for controlling BrowserOS — launch and automate the browser from the terminal or from AI coding agents like Claude Code and Gemini CLI.
-
-Communicates with the BrowserOS MCP server over JSON-RPC 2.0 / StreamableHTTP. All 53+ MCP tools are mapped to CLI commands.
-
-## Install
-
-### macOS / Linux
-
-```bash
-curl -fsSL https://cdn.browseros.com/cli/install.sh | bash
-```
-
-### Windows
-
-```powershell
-irm https://cdn.browseros.com/cli/install.ps1 | iex
-```
-
-### Build from Source
+## Setup
 
 Requires Go 1.25+.
 
 ```bash
-make            # Build binary
-make install    # Install to $GOPATH/bin
+# Build
+make
+
+# First run — configure server connection
+./browseros-cli init
 ```
 
-## Quick Start
+The `init` command prompts for your MCP server URL. Find it in:
+**BrowserOS → Settings → BrowserOS MCP → Server URL**
 
-```bash
-# If BrowserOS is not installed yet
-browseros-cli install                # downloads BrowserOS for your platform
+The port varies per installation (e.g., `http://127.0.0.1:9004/mcp`).
 
-# If BrowserOS is installed but not running
-browseros-cli launch                 # opens BrowserOS, waits for server
-
-# Configure the CLI (auto-discovers running BrowserOS)
-browseros-cli init --auto            # detects server URL and saves config
-
-# Verify connection
-browseros-cli health
-```
-
-### Other init modes
-
-```bash
-browseros-cli init <url>             # non-interactive — pass URL directly
-browseros-cli init                   # interactive — prompts for URL
-```
-
-Config is saved to `~/.config/browseros-cli/config.yaml`. The CLI also auto-discovers the server from `~/.browseros/server.json` (written by BrowserOS on startup).
-
-### CLI updates
-
-The CLI checks for a newer BrowserOS CLI release in the background about once per day and will suggest an update on a later run when one is available.
-
-```bash
-browseros-cli update         # check and apply the latest CLI release
-browseros-cli update --check # check only
-browseros-cli update --yes   # apply without prompting
-```
+Config is saved to `~/.config/browseros-cli/config.yaml`.
 
 ## Usage
 
@@ -110,12 +67,6 @@ browseros-cli history recent
 browseros-cli group list
 ```
 
-## Use as MCP Server
-
-BrowserOS exposes an MCP server that AI coding agents can connect to directly. The CLI is the easiest way to verify the connection and interact with tools from the terminal.
-
-To connect Claude Code, Gemini CLI, or any MCP client, see the [MCP setup guide](https://docs.browseros.com/features/use-with-claude-code).
-
 ## Global Flags
 
 | Flag | Env Var | Description |
@@ -126,9 +77,9 @@ To connect Claude Code, Gemini CLI, or any MCP client, see the [MCP setup guide]
 | `--debug` | `BOS_DEBUG=1` | Debug output |
 | `--timeout, -t` | | Request timeout (default: 2m) |
 
-Priority for server URL: `--server` flag > `BROWSEROS_URL` env > `~/.browseros/server.json` > config file
+Priority for server URL: `--server` flag > `BROWSEROS_URL` env > config file
 
-If no server URL is configured, the CLI exits with setup instructions pointing to `install`, `launch`, and `init`.
+If no server URL is configured, the CLI exits with setup instructions instead of assuming a localhost port.
 
 ## Testing
 
@@ -179,9 +130,7 @@ apps/cli/
 │   └── config.go       # Config file (~/.config/browseros-cli/config.yaml)
 ├── cmd/
 │   ├── root.go         # Root command, global flags
-│   ├── init.go         # Server URL configuration (URL arg, --auto, interactive)
-│   ├── install.go      # install (download BrowserOS for current platform)
-│   ├── launch.go       # launch (find and start BrowserOS, wait for server)
+│   ├── init.go         # Server URL configuration
 │   ├── open.go         # open (new_page / new_hidden_page)
 │   ├── nav.go          # nav, back, forward, reload
 │   ├── pages.go        # pages, active, close
@@ -214,8 +163,4 @@ The CLI communicates with BrowserOS via two HTTP POST requests per command:
 1. `initialize` — MCP handshake
 2. `tools/call` — execute the actual tool
 
-## Links
-
-- [Documentation](https://docs.browseros.com)
-- [MCP Setup Guide](https://docs.browseros.com/features/use-with-claude-code)
-- [Changelog](./CHANGELOG.md)
+All 54 MCP tools are mapped to CLI commands.

packages/browseros-agent/apps/cli/analytics/analytics.go

@@ -1,129 +0,0 @@
-package analytics
-
-import (
-	"crypto/rand"
-	"encoding/json"
-	"fmt"
-	"os"
-	"path/filepath"
-	"runtime"
-	"strings"
-	"time"
-
-	"browseros-cli/config"
-
-	"github.com/posthog/posthog-go"
-)
-
-var (
-	posthogAPIKey string // set via ldflags
-	posthogHost   = "https://us.i.posthog.com"
-)
-
-const eventPrefix = "browseros.cli."
-
-var svc *service
-
-type service struct {
-	client    posthog.Client
-	distinctID string
-}
-
-func Init(version string) {
-	if posthogAPIKey == "" {
-		return
-	}
-
-	distinctID := resolveDistinctID()
-	if distinctID == "" {
-		return
-	}
-
-	client, err := posthog.NewWithConfig(posthogAPIKey, posthog.Config{
-		Endpoint:        posthogHost,
-		BatchSize:       10,
-		ShutdownTimeout: 3 * time.Second,
-		DefaultEventProperties: posthog.NewProperties().
-			Set("cli_version", version).
-			Set("os", runtime.GOOS).
-			Set("arch", runtime.GOARCH),
-	})
-	if err != nil {
-		return
-	}
-
-	svc = &service{client: client, distinctID: distinctID}
-}
-
-func Track(command string, success bool, duration time.Duration) {
-	if svc == nil {
-		return
-	}
-	svc.client.Enqueue(posthog.Capture{
-		DistinctId: svc.distinctID,
-		Event:      eventPrefix + "command_executed",
-		Properties: posthog.NewProperties().
-			Set("command", command).
-			Set("success", success).
-			Set("duration_ms", duration.Milliseconds()).
-			Set("$process_person_profile", false),
-	})
-}
-
-func Close() {
-	if svc == nil {
-		return
-	}
-	svc.client.Close()
-	svc = nil
-}
-
-func resolveDistinctID() string {
-	if id := loadBrowserosID(); id != "" {
-		return id
-	}
-	return loadOrCreateInstallID()
-}
-
-func loadBrowserosID() string {
-	home, err := os.UserHomeDir()
-	if err != nil {
-		return ""
-	}
-	data, err := os.ReadFile(filepath.Join(home, ".browseros", "server.json"))
-	if err != nil {
-		return ""
-	}
-	var sc struct {
-		BrowserosID string `json:"browseros_id"`
-	}
-	if json.Unmarshal(data, &sc) != nil {
-		return ""
-	}
-	return sc.BrowserosID
-}
-
-func loadOrCreateInstallID() string {
-	dir := config.Dir()
-	idPath := filepath.Join(dir, "install_id")
-
-	data, err := os.ReadFile(idPath)
-	if err == nil {
-		if id := strings.TrimSpace(string(data)); id != "" {
-			return id
-		}
-	}
-
-	id := generateUUID()
-	os.MkdirAll(dir, 0755)
-	os.WriteFile(idPath, []byte(id), 0644)
-	return id
-}
-
-func generateUUID() string {
-	var b [16]byte
-	rand.Read(b[:])
-	b[6] = (b[6] & 0x0f) | 0x40 // version 4
-	b[8] = (b[8] & 0x3f) | 0x80 // variant 2
-	return fmt.Sprintf("%x-%x-%x-%x-%x", b[0:4], b[4:6], b[6:8], b[8:10], b[10:])
-}

packages/browseros-agent/apps/cli/analytics/analytics_test.go

@@ -1,132 +0,0 @@
-package analytics
-
-import (
-	"encoding/json"
-	"os"
-	"path/filepath"
-	"regexp"
-	"testing"
-	"time"
-)
-
-func TestGenerateUUID(t *testing.T) {
-	id := generateUUID()
-	uuidRe := regexp.MustCompile(`^[0-9a-f]{8}-[0-9a-f]{4}-4[0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$`)
-	if !uuidRe.MatchString(id) {
-		t.Errorf("generateUUID() = %q, does not match UUID v4 pattern", id)
-	}
-
-	id2 := generateUUID()
-	if id == id2 {
-		t.Error("generateUUID() returned the same value twice")
-	}
-}
-
-func TestLoadBrowserosID(t *testing.T) {
-	tmp := t.TempDir()
-	t.Setenv("HOME", tmp)
-
-	// No server.json → empty
-	if got := loadBrowserosID(); got != "" {
-		t.Errorf("loadBrowserosID() = %q, want empty", got)
-	}
-
-	// server.json without browseros_id → empty
-	dir := filepath.Join(tmp, ".browseros")
-	os.MkdirAll(dir, 0755)
-	data, _ := json.Marshal(map[string]any{"server_port": 9100, "url": "http://127.0.0.1:9100"})
-	os.WriteFile(filepath.Join(dir, "server.json"), data, 0644)
-
-	if got := loadBrowserosID(); got != "" {
-		t.Errorf("loadBrowserosID() = %q, want empty (no browseros_id field)", got)
-	}
-
-	// server.json with browseros_id → returns it
-	data, _ = json.Marshal(map[string]any{
-		"server_port":  9100,
-		"url":          "http://127.0.0.1:9100",
-		"browseros_id": "test-uuid-1234",
-	})
-	os.WriteFile(filepath.Join(dir, "server.json"), data, 0644)
-
-	if got := loadBrowserosID(); got != "test-uuid-1234" {
-		t.Errorf("loadBrowserosID() = %q, want %q", got, "test-uuid-1234")
-	}
-}
-
-func TestLoadOrCreateInstallID(t *testing.T) {
-	tmp := t.TempDir()
-	configDir := filepath.Join(tmp, "browseros-cli")
-	t.Setenv("XDG_CONFIG_HOME", tmp)
-
-	// First call creates the file
-	id := loadOrCreateInstallID()
-	if id == "" {
-		t.Fatal("loadOrCreateInstallID() returned empty string")
-	}
-
-	// File was persisted
-	data, err := os.ReadFile(filepath.Join(configDir, "install_id"))
-	if err != nil {
-		t.Fatalf("install_id file not created: %v", err)
-	}
-	if string(data) != id {
-		t.Errorf("persisted id = %q, want %q", string(data), id)
-	}
-
-	// Second call returns the same ID
-	id2 := loadOrCreateInstallID()
-	if id2 != id {
-		t.Errorf("loadOrCreateInstallID() = %q, want stable %q", id2, id)
-	}
-}
-
-func TestResolveDistinctID_PrefersBrowserosID(t *testing.T) {
-	tmp := t.TempDir()
-	t.Setenv("HOME", tmp)
-	t.Setenv("XDG_CONFIG_HOME", tmp)
-
-	// Write server.json with browseros_id
-	dir := filepath.Join(tmp, ".browseros")
-	os.MkdirAll(dir, 0755)
-	data, _ := json.Marshal(map[string]any{"browseros_id": "server-uuid"})
-	os.WriteFile(filepath.Join(dir, "server.json"), data, 0644)
-
-	got := resolveDistinctID()
-	if got != "server-uuid" {
-		t.Errorf("resolveDistinctID() = %q, want %q", got, "server-uuid")
-	}
-}
-
-func TestResolveDistinctID_FallsBackToInstallID(t *testing.T) {
-	tmp := t.TempDir()
-	t.Setenv("HOME", tmp)
-	t.Setenv("XDG_CONFIG_HOME", tmp)
-
-	// No server.json → should generate install_id
-	got := resolveDistinctID()
-	if got == "" {
-		t.Error("resolveDistinctID() returned empty string")
-	}
-}
-
-func TestInitNoopsWithoutAPIKey(t *testing.T) {
-	old := posthogAPIKey
-	posthogAPIKey = ""
-	defer func() { posthogAPIKey = old }()
-
-	Init("1.0.0")
-	if svc != nil {
-		t.Error("Init() created service without API key")
-	}
-}
-
-func TestTrackAndCloseNoopWithoutInit(t *testing.T) {
-	old := svc
-	svc = nil
-	defer func() { svc = old }()
-
-	// Should not panic
-	Track("test", true, time.Second)
-	Close()
-}

packages/browseros-agent/apps/cli/cmd/health.go

@@ -49,7 +49,7 @@ func init() {
 	statusCmd := &cobra.Command{
 		Use:         "status",
 		Annotations: map[string]string{"group": "Setup:"},
-		Short:       "Check BrowserOS runtime status",
+		Short:       "Check extension connection status",
 		Args:        cobra.NoArgs,
 		Run: func(cmd *cobra.Command, args []string) {
 			c := newClient()
@@ -64,12 +64,12 @@ func init() {
 			green := color.New(color.FgGreen).SprintFunc()
 			red := color.New(color.FgRed).SprintFunc()
 
-			cdp := data["cdpConnected"]
-			cdpStr := red("disconnected")
-			if b, ok := cdp.(bool); ok && b {
-				cdpStr = green("connected")
+			ext := data["extensionConnected"]
+			extStr := red("disconnected")
+			if b, ok := ext.(bool); ok && b {
+				extStr = green("connected")
 			}
-			fmt.Printf("Browser: %s\n", cdpStr)
+			fmt.Printf("Extension: %s\n", extStr)
 		},
 	}
 

packages/browseros-agent/apps/cli/cmd/init.go

@@ -25,17 +25,13 @@ func init() {
 		Long: `Set up the CLI by providing the MCP server URL from BrowserOS.
 
 Open BrowserOS → Settings → BrowserOS MCP to find your Server URL.
-The URL looks like: http://127.0.0.1:9000/mcp
+The URL looks like: http://127.0.0.1:9004/mcp
 
 The port varies per installation, so this step is required on first use.
 Run again if your port changes.
 
-You can provide the full URL or just the port number:
-  browseros-cli init http://127.0.0.1:9000/mcp
-  browseros-cli init 9000
-
 Three modes:
-  browseros-cli init <url>    Non-interactive (full URL or port number)
+  browseros-cli init <url>    Non-interactive, use the provided URL
   browseros-cli init --auto   Auto-discover from ~/.browseros/server.json
   browseros-cli init          Interactive prompt`,
 		Annotations: map[string]string{"group": "Setup:"},
@@ -69,14 +65,13 @@ Three modes:
 				bold.Println("BrowserOS CLI Setup")
 				fmt.Println()
 				fmt.Println("Open BrowserOS → Settings → BrowserOS MCP")
-				fmt.Println("Copy the Server URL or port number shown there.")
+				fmt.Println("Copy the Server URL shown there.")
 				fmt.Println()
-				dim.Println("Examples:  http://127.0.0.1:9000/mcp")
-				dim.Println("           9000")
+				dim.Println("It looks like: http://127.0.0.1:9004/mcp")
 				fmt.Println()
 
 				reader := bufio.NewReader(os.Stdin)
-				fmt.Print("Server URL or port: ")
+				fmt.Print("Server URL: ")
 				line, err := reader.ReadString('\n')
 				if err != nil {
 					output.Error("failed to read input", 1)

packages/browseros-agent/apps/cli/cmd/install.go

@@ -148,7 +148,7 @@ func runPostInstall(path string, deb bool, dim *color.Color) {
 // installMacOS mounts the DMG and copies BrowserOS.app to /Applications.
 func installMacOS(dmgPath string, dim *color.Color) {
 	fmt.Println("Mounting disk image...")
-	mountOut, err := exec.Command("hdiutil", "attach", dmgPath, "-nobrowse").Output()
+	mountOut, err := exec.Command("hdiutil", "attach", dmgPath, "-nobrowse", "-quiet").Output()
 	if err != nil {
 		dim.Println("Could not mount DMG automatically.")
 		dim.Printf("  Open it manually: open %s\n", dmgPath)

packages/browseros-agent/apps/cli/cmd/root.go

@@ -1,7 +1,6 @@
 package cmd
 
 import (
-	"context"
 	"encoding/json"
 	"fmt"
 	"os"
@@ -10,11 +9,9 @@ import (
 	"strings"
 	"time"
 
-	"browseros-cli/analytics"
 	"browseros-cli/config"
 	"browseros-cli/mcp"
 	"browseros-cli/output"
-	"browseros-cli/update"
 
 	"github.com/fatih/color"
 	"github.com/spf13/cobra"
@@ -30,11 +27,8 @@ var (
 	version   = "dev"
 )
 
-const automaticUpdateDrainTimeout = 150 * time.Millisecond
-
 func SetVersion(v string) {
 	version = v
-	rootCmd.Version = v
 }
 
 var (
@@ -119,40 +113,11 @@ var rootCmd = &cobra.Command{
 }
 
 func Execute() {
-	automaticUpdater := newAutomaticUpdateManager(os.Args[1:])
-	automaticNotice := ""
-	var automaticCheckDone <-chan struct{}
-	if automaticUpdater != nil {
-		automaticNotice = automaticUpdater.CachedNotice()
-		automaticCheckDone = automaticUpdater.StartBackgroundCheck(context.Background())
-	}
-
-	analytics.Init(version)
-	start := time.Now()
-
-	err := rootCmd.Execute()
-
-	if automaticNotice != "" && err == nil {
-		fmt.Fprintln(os.Stderr, automaticNotice)
-	}
-	drainAutomaticUpdateCheck(automaticCheckDone)
-
-	analytics.Track(commandName(os.Args[1:]), err == nil, time.Since(start))
-	analytics.Close()
-
-	if err != nil {
+	if err := rootCmd.Execute(); err != nil {
 		os.Exit(1)
 	}
 }
 
-func commandName(args []string) string {
-	cmd, _, err := rootCmd.Find(args)
-	if err != nil || cmd == rootCmd {
-		return "unknown"
-	}
-	return cmd.CommandPath()
-}
-
 func init() {
 	cobra.AddTemplateFunc("helpHeader", helpHeader)
 	cobra.AddTemplateFunc("helpCmdCol", helpCmdCol)
@@ -201,93 +166,6 @@ func envBool(key string) bool {
 	return v == "1" || v == "true"
 }
 
-func newAutomaticUpdateManager(args []string) *update.Manager {
-	if shouldSkipAutomaticUpdates(args) {
-		return nil
-	}
-
-	return update.NewManager(update.Options{
-		CurrentVersion: version,
-		JSONOutput:     requestedBoolFlag(args, "--json", jsonOut),
-		Debug:          requestedBoolFlag(args, "--debug", debug),
-		Automatic:      true,
-	})
-}
-
-func shouldSkipAutomaticUpdates(args []string) bool {
-	if hasHelpFlag(args) || requestedBoolFlag(args, "--version", false) {
-		return true
-	}
-
-	switch primaryCommand(args) {
-	case "help", "completion", "update", "self-update", "upgrade":
-		return true
-	default:
-		return false
-	}
-}
-
-func hasHelpFlag(args []string) bool {
-	if requestedBoolFlag(args, "--help", false) {
-		return true
-	}
-
-	for _, arg := range args {
-		if arg == "-h" {
-			return true
-		}
-	}
-
-	return false
-}
-
-func primaryCommand(args []string) string {
-	for _, arg := range args {
-		if strings.HasPrefix(arg, "-") {
-			continue
-		}
-		return arg
-	}
-	return ""
-}
-
-func requestedBoolFlag(args []string, flagName string, current bool) bool {
-	if current {
-		return true
-	}
-
-	prefix := flagName + "="
-	for _, arg := range args {
-		if arg == flagName {
-			return true
-		}
-		if strings.HasPrefix(arg, prefix) {
-			value, err := strconv.ParseBool(strings.TrimPrefix(arg, prefix))
-			return err == nil && value
-		}
-	}
-
-	return false
-}
-
-func drainAutomaticUpdateCheck(done <-chan struct{}) {
-	drainAutomaticUpdateCheckWithTimeout(done, automaticUpdateDrainTimeout)
-}
-
-func drainAutomaticUpdateCheckWithTimeout(done <-chan struct{}, timeout time.Duration) {
-	if done == nil {
-		return
-	}
-
-	timer := time.NewTimer(timeout)
-	defer timer.Stop()
-
-	select {
-	case <-done:
-	case <-timer.C:
-	}
-}
-
 func defaultServerURL() string {
 	// 1. Explicit env var always wins
 	if env := normalizeServerURL(os.Getenv("BROWSEROS_URL")); env != "" {
@@ -339,27 +217,10 @@ func loadBrowserosServerURL() string {
 
 func normalizeServerURL(raw string) string {
 	normalized := strings.TrimSpace(raw)
-
-	if isPortOnly(normalized) {
-		normalized = "http://127.0.0.1:" + normalized
-	}
-
 	normalized = strings.TrimSuffix(normalized, "/mcp")
 	return strings.TrimSuffix(normalized, "/")
 }
 
-func isPortOnly(s string) bool {
-	if s == "" {
-		return false
-	}
-	for _, c := range s {
-		if c < '0' || c > '9' {
-			return false
-		}
-	}
-	return true
-}
-
 func validateServerURL(raw string) (string, error) {
 	baseURL := normalizeServerURL(raw)
 	if baseURL != "" {

packages/browseros-agent/apps/cli/cmd/root_test.go

@@ -1,146 +0,0 @@
-package cmd
-
-import (
-	"testing"
-	"time"
-)
-
-func TestSetVersionUpdatesRootCommand(t *testing.T) {
-	originalVersion := version
-	originalRootVersion := rootCmd.Version
-	t.Cleanup(func() {
-		version = originalVersion
-		rootCmd.Version = originalRootVersion
-	})
-
-	SetVersion("1.2.3")
-
-	if version != "1.2.3" {
-		t.Fatalf("version = %q, want %q", version, "1.2.3")
-	}
-	if rootCmd.Version != "1.2.3" {
-		t.Fatalf("rootCmd.Version = %q, want %q", rootCmd.Version, "1.2.3")
-	}
-}
-
-func TestCommandName(t *testing.T) {
-	tests := []struct {
-		name string
-		args []string
-		want string
-	}{
-		{"empty args", nil, "unknown"},
-		{"known command", []string{"health"}, "browseros-cli health"},
-		{"unknown command", []string{"nonexistent"}, "unknown"},
-		{"subcommand", []string{"bookmark", "search"}, "browseros-cli bookmark search"},
-		{"known with extra args", []string{"snap", "--enhanced"}, "browseros-cli snap"},
-	}
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			got := commandName(tt.args)
-			if got != tt.want {
-				t.Errorf("commandName(%v) = %q, want %q", tt.args, got, tt.want)
-			}
-		})
-	}
-}
-
-func TestPrimaryCommand(t *testing.T) {
-	tests := []struct {
-		name string
-		args []string
-		want string
-	}{
-		{"empty", nil, ""},
-		{"root flag then command", []string{"--json", "update"}, "update"},
-		{"subcommand", []string{"bookmark", "update"}, "bookmark"},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			if got := primaryCommand(tt.args); got != tt.want {
-				t.Fatalf("primaryCommand(%v) = %q, want %q", tt.args, got, tt.want)
-			}
-		})
-	}
-}
-
-func TestRequestedBoolFlag(t *testing.T) {
-	if !requestedBoolFlag([]string{"--json"}, "--json", false) {
-		t.Fatal("requestedBoolFlag() = false, want true")
-	}
-	if !requestedBoolFlag([]string{"--debug=true"}, "--debug", false) {
-		t.Fatal("requestedBoolFlag() with assignment = false, want true")
-	}
-	if requestedBoolFlag([]string{"--debug=false"}, "--debug", false) {
-		t.Fatal("requestedBoolFlag() with false assignment = true, want false")
-	}
-}
-
-func TestShouldSkipAutomaticUpdates(t *testing.T) {
-	tests := []struct {
-		name string
-		args []string
-		want bool
-	}{
-		{"short help flag", []string{"-h"}, true},
-		{"help flag", []string{"--help"}, true},
-		{"version flag", []string{"--version"}, true},
-		{"update command", []string{"update"}, true},
-		{"bookmark update subcommand", []string{"bookmark", "update"}, false},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			if got := shouldSkipAutomaticUpdates(tt.args); got != tt.want {
-				t.Fatalf("shouldSkipAutomaticUpdates(%v) = %t, want %t", tt.args, got, tt.want)
-			}
-		})
-	}
-}
-
-func TestDrainAutomaticUpdateCheckWithTimeoutWaitsForCompletion(t *testing.T) {
-	done := make(chan struct{})
-	returned := make(chan struct{})
-
-	go func() {
-		drainAutomaticUpdateCheckWithTimeout(done, time.Second)
-		close(returned)
-	}()
-
-	select {
-	case <-returned:
-		t.Fatal("drainAutomaticUpdateCheckWithTimeout() returned before check completed")
-	case <-time.After(10 * time.Millisecond):
-	}
-
-	close(done)
-
-	select {
-	case <-returned:
-	case <-time.After(100 * time.Millisecond):
-		t.Fatal("drainAutomaticUpdateCheckWithTimeout() did not return after check completed")
-	}
-}
-
-func TestDrainAutomaticUpdateCheckWithTimeoutStopsWaiting(t *testing.T) {
-	done := make(chan struct{})
-	returned := make(chan struct{})
-
-	go func() {
-		drainAutomaticUpdateCheckWithTimeout(done, 20*time.Millisecond)
-		close(returned)
-	}()
-
-	select {
-	case <-returned:
-		t.Fatal("drainAutomaticUpdateCheckWithTimeout() returned before timeout elapsed")
-	case <-time.After(5 * time.Millisecond):
-	}
-
-	select {
-	case <-returned:
-	case <-time.After(100 * time.Millisecond):
-		t.Fatal("drainAutomaticUpdateCheckWithTimeout() did not return after timeout")
-	}
-}

packages/browseros-agent/apps/cli/cmd/update.go

@@ -1,179 +0,0 @@
-package cmd
-
-import (
-	"bufio"
-	"context"
-	"fmt"
-	"io"
-	"os"
-	"strings"
-
-	"browseros-cli/output"
-	"browseros-cli/update"
-
-	"github.com/spf13/cobra"
-)
-
-type updateManager interface {
-	CheckNow(context.Context) (*update.CheckResult, error)
-	Apply(context.Context, *update.CheckResult) error
-}
-
-type updateOutcome struct {
-	result   *update.CheckResult
-	applied  bool
-	canceled bool
-}
-
-func init() {
-	cmd := &cobra.Command{
-		Use:         "update",
-		Aliases:     []string{"self-update", "upgrade"},
-		Annotations: map[string]string{"group": "Setup:"},
-		Short:       "Check for and apply CLI updates",
-		Args:        cobra.NoArgs,
-		Run: func(cmd *cobra.Command, args []string) {
-			checkOnly, _ := cmd.Flags().GetBool("check")
-			yes, _ := cmd.Flags().GetBool("yes")
-
-			manager := update.NewManager(update.Options{
-				CurrentVersion: version,
-				JSONOutput:     jsonOut,
-				Debug:          debug,
-				Automatic:      false,
-			})
-			outcome, err := runUpdateCommand(
-				cmd.Context(),
-				manager,
-				checkOnly,
-				yes,
-				stdinIsInteractive(os.Stdin),
-				os.Stdin,
-				os.Stderr,
-			)
-			if err != nil {
-				output.Error(err.Error(), 1)
-			}
-			printUpdateOutcome(outcome)
-		},
-	}
-
-	cmd.Flags().Bool("check", false, "Check for updates without applying them")
-	cmd.Flags().Bool("yes", false, "Apply update without prompting")
-
-	rootCmd.AddCommand(cmd)
-}
-
-func runUpdateCommand(
-	ctx context.Context,
-	manager updateManager,
-	checkOnly bool,
-	yes bool,
-	interactive bool,
-	stdin io.Reader,
-	stderr io.Writer,
-) (*updateOutcome, error) {
-	result, err := manager.CheckNow(ctx)
-	if err != nil {
-		return nil, err
-	}
-
-	outcome := &updateOutcome{result: result}
-	if checkOnly || !result.UpdateAvailable {
-		return outcome, nil
-	}
-	if !yes {
-		if !interactive {
-			return nil, fmt.Errorf("update requires confirmation; rerun with --yes")
-		}
-
-		confirmed, err := confirmUpdate(stdin, stderr, result)
-		if err != nil {
-			return nil, err
-		}
-		if !confirmed {
-			outcome.canceled = true
-			return outcome, nil
-		}
-	}
-	if err := manager.Apply(ctx, result); err != nil {
-		return nil, err
-	}
-	outcome.applied = true
-
-	return outcome, nil
-}
-
-func printUpdateOutcome(outcome *updateOutcome) {
-	if jsonOut {
-		output.JSONRaw(updateOutcomePayload(outcome))
-		return
-	}
-
-	switch {
-	case outcome.applied:
-		fmt.Printf("Updated browseros-cli to v%s\n", outcome.result.LatestVersion)
-	case outcome.canceled:
-		fmt.Println("Update canceled.")
-	case outcome.result.UpdateAvailable:
-		fmt.Println(update.FormatNotice(outcome.result.CurrentVersion, outcome.result.LatestVersion))
-	case outcome.result != nil:
-		fmt.Printf("browseros-cli is up to date (v%s)\n", outcome.result.CurrentVersion)
-	}
-}
-
-func updateOutcomePayload(outcome *updateOutcome) map[string]any {
-	payload := map[string]any{
-		"applied": outcome.applied,
-	}
-	if outcome.canceled {
-		payload["canceled"] = true
-	}
-	if outcome.result == nil {
-		return payload
-	}
-
-	payload["currentVersion"] = outcome.result.CurrentVersion
-	payload["latestVersion"] = outcome.result.LatestVersion
-	payload["updateAvailable"] = outcome.result.UpdateAvailable
-	if outcome.result.Asset != nil {
-		payload["asset"] = map[string]any{
-			"filename":      outcome.result.Asset.Filename,
-			"url":           outcome.result.Asset.URL,
-			"archiveFormat": outcome.result.Asset.ArchiveFormat,
-		}
-	}
-
-	return payload
-}
-
-func confirmUpdate(
-	stdin io.Reader,
-	stderr io.Writer,
-	result *update.CheckResult,
-) (bool, error) {
-	if _, err := fmt.Fprintf(
-		stderr,
-		"Install browseros-cli v%s over v%s? [y/N]: ",
-		result.LatestVersion,
-		result.CurrentVersion,
-	); err != nil {
-		return false, err
-	}
-
-	line, err := bufio.NewReader(stdin).ReadString('\n')
-	if err != nil && err != io.EOF {
-		return false, err
-	}
-	answer := strings.ToLower(strings.TrimSpace(line))
-
-	return answer == "y" || answer == "yes", nil
-}
-
-func stdinIsInteractive(file *os.File) bool {
-	info, err := file.Stat()
-	if err != nil {
-		return false
-	}
-	return info.Mode()&os.ModeCharDevice != 0
-}

packages/browseros-agent/apps/cli/cmd/update_test.go

@@ -1,176 +0,0 @@
-package cmd
-
-import (
-	"bytes"
-	"context"
-	"errors"
-	"net/http"
-	"net/http/httptest"
-	"runtime"
-	"testing"
-
-	"browseros-cli/update"
-)
-
-func TestRunUpdateCommandCheckOnly(t *testing.T) {
-	configRoot := t.TempDir()
-	t.Setenv("XDG_CONFIG_HOME", configRoot)
-
-	manager := newTestUpdateManager(t)
-	outcome, err := runUpdateCommand(
-		context.Background(),
-		manager,
-		true,
-		false,
-		false,
-		bytes.NewBufferString(""),
-		&bytes.Buffer{},
-	)
-	if err != nil {
-		t.Fatalf("runUpdateCommand() error = %v", err)
-	}
-	if outcome.applied {
-		t.Fatal("runUpdateCommand() applied = true, want false")
-	}
-	if !outcome.result.UpdateAvailable {
-		t.Fatal("runUpdateCommand() UpdateAvailable = false, want true")
-	}
-}
-
-func TestRunUpdateCommandRequiresYesWithoutTTY(t *testing.T) {
-	configRoot := t.TempDir()
-	t.Setenv("XDG_CONFIG_HOME", configRoot)
-
-	_, err := runUpdateCommand(
-		context.Background(),
-		newTestUpdateManager(t),
-		false,
-		false,
-		false,
-		bytes.NewBufferString(""),
-		&bytes.Buffer{},
-	)
-	if err == nil {
-		t.Fatal("runUpdateCommand() error = nil, want confirmation error")
-	}
-}
-
-func TestRunUpdateCommandCancel(t *testing.T) {
-	configRoot := t.TempDir()
-	t.Setenv("XDG_CONFIG_HOME", configRoot)
-
-	stderr := &bytes.Buffer{}
-	outcome, err := runUpdateCommand(
-		context.Background(),
-		newTestUpdateManager(t),
-		false,
-		false,
-		true,
-		bytes.NewBufferString("n\n"),
-		stderr,
-	)
-	if err != nil {
-		t.Fatalf("runUpdateCommand() error = %v", err)
-	}
-	if !outcome.canceled {
-		t.Fatal("runUpdateCommand() canceled = false, want true")
-	}
-	if stderr.Len() == 0 {
-		t.Fatal("confirm prompt was not written to stderr")
-	}
-}
-
-func TestRunUpdateCommandYesAppliesWithoutPrompt(t *testing.T) {
-	manager := &fakeUpdateManager{
-		result: &update.CheckResult{
-			CurrentVersion:  "1.0.0",
-			LatestVersion:   "9.9.9",
-			UpdateAvailable: true,
-			Asset: &update.Asset{
-				Filename:      "browseros-cli_9.9.9_test.tar.gz",
-				URL:           "https://cdn.example.com/cli/v9.9.9/browseros-cli_9.9.9_test.tar.gz",
-				ArchiveFormat: "tar.gz",
-				SHA256:        "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
-			},
-		},
-	}
-	stderr := &bytes.Buffer{}
-
-	outcome, err := runUpdateCommand(
-		context.Background(),
-		manager,
-		false,
-		true,
-		false,
-		bytes.NewBufferString(""),
-		stderr,
-	)
-	if err != nil {
-		t.Fatalf("runUpdateCommand() error = %v", err)
-	}
-	if !outcome.applied {
-		t.Fatal("runUpdateCommand() applied = false, want true")
-	}
-	if manager.applyCalls != 1 {
-		t.Fatalf("Apply() calls = %d, want 1", manager.applyCalls)
-	}
-	if stderr.Len() != 0 {
-		t.Fatal("prompt was written despite --yes")
-	}
-}
-
-type fakeUpdateManager struct {
-	result     *update.CheckResult
-	checkErr   error
-	applyErr   error
-	applyCalls int
-}
-
-func (m *fakeUpdateManager) CheckNow(context.Context) (*update.CheckResult, error) {
-	if m.checkErr != nil {
-		return nil, m.checkErr
-	}
-	if m.result == nil {
-		return nil, errors.New("missing check result")
-	}
-	return m.result, nil
-}
-
-func (m *fakeUpdateManager) Apply(context.Context, *update.CheckResult) error {
-	m.applyCalls++
-	return m.applyErr
-}
-
-func newTestUpdateManager(t *testing.T) *update.Manager {
-	t.Helper()
-
-	key, err := update.PlatformKey(runtime.GOOS, runtime.GOARCH)
-	if err != nil {
-		t.Fatalf("PlatformKey() error = %v", err)
-	}
-
-	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "application/json")
-		_, _ = w.Write([]byte(`{
-			"version":"9.9.9",
-			"published_at":"2026-03-27T19:00:00Z",
-			"tag":"browseros-cli-v9.9.9",
-			"assets":{
-				"` + key + `":{
-					"filename":"browseros-cli_9.9.9_test.tar.gz",
-					"url":"https://cdn.example.com/cli/v9.9.9/browseros-cli_9.9.9_test.tar.gz",
-					"archive_format":"tar.gz",
-					"sha256":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
-				}
-			}
-		}`))
-	}))
-	t.Cleanup(server.Close)
-
-	return update.NewManager(update.Options{
-		CurrentVersion: "1.0.0",
-		ManifestURL:    server.URL,
-		Automatic:      false,
-		HTTPClient:     server.Client(),
-	})
-}

packages/browseros-agent/apps/cli/go.mod

@@ -4,28 +4,20 @@ go 1.25.7
 
 require (
 	github.com/fatih/color v1.18.0
-	github.com/minio/selfupdate v0.6.0
-	github.com/modelcontextprotocol/go-sdk v1.4.0
-	github.com/posthog/posthog-go v1.11.2
 	github.com/spf13/cobra v1.10.2
-	golang.org/x/mod v0.34.0
-	gopkg.in/yaml.v3 v3.0.1
 )
 
 require (
-	aead.dev/minisign v0.2.0 // indirect
-	github.com/goccy/go-json v0.10.5 // indirect
 	github.com/google/jsonschema-go v0.4.2 // indirect
-	github.com/google/uuid v1.6.0 // indirect
-	github.com/hashicorp/golang-lru/v2 v2.0.7 // indirect
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
 	github.com/mattn/go-colorable v0.1.13 // indirect
 	github.com/mattn/go-isatty v0.0.20 // indirect
+	github.com/modelcontextprotocol/go-sdk v1.4.0 // indirect
 	github.com/segmentio/asm v1.1.3 // indirect
 	github.com/segmentio/encoding v0.5.3 // indirect
 	github.com/spf13/pflag v1.0.9 // indirect
 	github.com/yosida95/uritemplate/v3 v3.0.2 // indirect
-	golang.org/x/crypto v0.0.0-20211209193657-4570a0811e8b // indirect
 	golang.org/x/oauth2 v0.34.0 // indirect
 	golang.org/x/sys v0.40.0 // indirect
+	gopkg.in/yaml.v3 v3.0.1 // indirect
 )

packages/browseros-agent/apps/cli/go.sum

@@ -1,22 +1,8 @@
-aead.dev/minisign v0.2.0 h1:kAWrq/hBRu4AARY6AlciO83xhNnW9UaC8YipS2uhLPk=
-aead.dev/minisign v0.2.0/go.mod h1:zdq6LdSd9TbuSxchxwhpA9zEb9YXcVGoE8JakuiGaIQ=
 github.com/cpuguy83/go-md2man/v2 v2.0.6/go.mod h1:oOW0eioCTA6cOiMLiUPZOpcVxMig6NIQQ7OS05n1F4g=
-github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
-github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/fatih/color v1.18.0 h1:S8gINlzdQ840/4pfAwic/ZE0djQEH3wM94VfqLTZcOM=
 github.com/fatih/color v1.18.0/go.mod h1:4FelSpRwEGDpQ12mAdzqdOukCy4u8WUtOY6lkT/6HfU=
-github.com/goccy/go-json v0.10.5 h1:Fq85nIqj+gXn/S5ahsiTlK3TmC85qgirsdTP/+DeaC4=
-github.com/goccy/go-json v0.10.5/go.mod h1:oq7eo15ShAhp70Anwd5lgX2pLfOS3QCiwU/PULtXL6M=
-github.com/golang-jwt/jwt/v5 v5.3.0 h1:pv4AsKCKKZuqlgs5sUmn4x8UlGa0kEVt/puTpKx9vvo=
-github.com/golang-jwt/jwt/v5 v5.3.0/go.mod h1:fxCRLWMO43lRc8nhHWY6LGqRcf+1gQWArsqaEUEa5bE=
-github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
-github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
 github.com/google/jsonschema-go v0.4.2 h1:tmrUohrwoLZZS/P3x7ex0WAVknEkBZM46iALbcqoRA8=
 github.com/google/jsonschema-go v0.4.2/go.mod h1:r5quNTdLOYEz95Ru18zA0ydNbBuYoo9tgaYcxEYhJVE=
-github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
-github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
-github.com/hashicorp/golang-lru/v2 v2.0.7 h1:a+bsQ5rvGLjzHuww6tVxozPZFVghXaHOwFs4luLUK2k=
-github.com/hashicorp/golang-lru/v2 v2.0.7/go.mod h1:QeFd9opnmA6QUJc5vARoKUSoFhyfM2/ZepoAG6RGpeM=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
 github.com/mattn/go-colorable v0.1.13 h1:fFA4WZxdEF4tXPZVKMLwD8oUnCTTo08duU7wxecdEvA=
@@ -24,14 +10,8 @@ github.com/mattn/go-colorable v0.1.13/go.mod h1:7S9/ev0klgBDR4GtXTXX8a3vIGJpMovk
 github.com/mattn/go-isatty v0.0.16/go.mod h1:kYGgaQfpe5nmfYZH+SKPsOc2e4SrIfOl2e/yFXSvRLM=
 github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
 github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
-github.com/minio/selfupdate v0.6.0 h1:i76PgT0K5xO9+hjzKcacQtO7+MjJ4JKA8Ak8XQ9DDwU=
-github.com/minio/selfupdate v0.6.0/go.mod h1:bO02GTIPCMQFTEvE5h4DjYB58bCoZ35XLeBf0buTDdM=
 github.com/modelcontextprotocol/go-sdk v1.4.0 h1:u0kr8lbJc1oBcawK7Df+/ajNMpIDFE41OEPxdeTLOn8=
 github.com/modelcontextprotocol/go-sdk v1.4.0/go.mod h1:Nxc2n+n/GdCebUaqCOhTetptS17SXXNu9IfNTaLDi1E=
-github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
-github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
-github.com/posthog/posthog-go v1.11.2 h1:ApKTtOhIeWhUBc4ByO+mlbg2o0iZaEGJnJHX2QDnn5Q=
-github.com/posthog/posthog-go v1.11.2/go.mod h1:xsVOW9YImilUcazwPNEq4PJDqEZf2KeCS758zXjwkPg=
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
 github.com/segmentio/asm v1.1.3 h1:WM03sfUOENvvKexOLp+pCqgb/WDjsi7EK8gIsICtzhc=
 github.com/segmentio/asm v1.1.3/go.mod h1:Ld3L4ZXGNcSLRg4JBsZ3//1+f/TjYl0Mzen/DQy1EJg=
@@ -41,39 +21,17 @@ github.com/spf13/cobra v1.10.2 h1:DMTTonx5m65Ic0GOoRY2c16WCbHxOOw6xxezuLaBpcU=
 github.com/spf13/cobra v1.10.2/go.mod h1:7C1pvHqHw5A4vrJfjNwvOdzYu0Gml16OCs2GRiTUUS4=
 github.com/spf13/pflag v1.0.9 h1:9exaQaMOCwffKiiiYk6/BndUBv+iRViNW+4lEMi0PvY=
 github.com/spf13/pflag v1.0.9/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
-github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu7U=
-github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
 github.com/yosida95/uritemplate/v3 v3.0.2 h1:Ed3Oyj9yrmi9087+NczuL5BwkIc4wvTb5zIM+UJPGz4=
 github.com/yosida95/uritemplate/v3 v3.0.2/go.mod h1:ILOh0sOhIJR3+L/8afwt/kE++YT040gmv5BQTMR2HP4=
 go.yaml.in/yaml/v3 v3.0.4/go.mod h1:DhzuOOF2ATzADvBadXxruRBLzYTpT36CKvDb3+aBEFg=
-golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
-golang.org/x/crypto v0.0.0-20210220033148-5ea612d1eb83/go.mod h1:jdWPYTVW3xRLrWPugEBEK3UY2ZEsg3UU495nc5E+M+I=
-golang.org/x/crypto v0.0.0-20211209193657-4570a0811e8b h1:QAqMVf3pSa6eeTsuklijukjXBlj7Es2QQplab+/RbQ4=
-golang.org/x/crypto v0.0.0-20211209193657-4570a0811e8b/go.mod h1:IxCIyHEi3zRg3s0A5j5BB6A9Jmi73HwBIUl50j+osU4=
-golang.org/x/mod v0.34.0 h1:xIHgNUUnW6sYkcM5Jleh05DvLOtwc6RitGHbDk4akRI=
-golang.org/x/mod v0.34.0/go.mod h1:ykgH52iCZe79kzLLMhyCUzhMci+nQj+0XkbXpNYtVjY=
-golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
-golang.org/x/net v0.0.0-20211112202133-69e39bad7dc2/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
 golang.org/x/oauth2 v0.34.0 h1:hqK/t4AKgbqWkdkcAeI8XLmbK+4m4G5YeQRrmiotGlw=
 golang.org/x/oauth2 v0.34.0/go.mod h1:lzm5WQJQwKZ3nwavOZ3IS5Aulzxi68dUSgRHujetwEA=
-golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
-golang.org/x/sys v0.0.0-20191026070338-33540a1f6037/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210228012217-479acdf4ea46/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210615035016-665e8c7367d1/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.0.0-20220811171246-fbc7d0a398ab/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.25.0 h1:r+8e+loiHxRqhXVl6ML1nO3l1+oFoWbnlu2Ehimmi34=
+golang.org/x/sys v0.25.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
 golang.org/x/sys v0.40.0 h1:DBZZqJ2Rkml6QMQsZywtnjnnGvHza6BTfYFWY9kjEWQ=
 golang.org/x/sys v0.40.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks=
-golang.org/x/term v0.0.0-20201117132131-f5c789dd3221/go.mod h1:Nr5EML6q2oocZ2LXRh80K7BxOlk5/8JxuGnuhpl+muw=
-golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
-golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
-golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
-golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
-golang.org/x/tools v0.42.0 h1:uNgphsn75Tdz5Ji2q36v/nsFSfR/9BRFvqhGBaJGd5k=
-golang.org/x/tools v0.42.0/go.mod h1:Ma6lCIwGZvHK6XtgbswSoWroEkhugApmsXyrUmBhfr0=
-gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405 h1:yhCVgyC4o1eVCa2tZl7eS0r+SDo693bJlVdllGtEeKM=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
 gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

packages/browseros-agent/apps/cli/npm/.npmignore

@@ -1,2 +0,0 @@
-.binary/
-node_modules/

packages/browseros-agent/apps/cli/npm/README.md

@@ -1,81 +0,0 @@
-# browseros-cli
-
-Command-line interface for controlling BrowserOS -- launch and automate the browser from the terminal.
-
-## Installation
-
-**Zero install (recommended):**
-
-```bash
-npx browseros-cli --help
-```
-
-**Global install:**
-
-```bash
-npm install -g browseros-cli
-```
-
-**Shell script fallback:**
-
-```bash
-curl -fsSL https://cdn.browseros.com/cli/install.sh | bash
-```
-
-## Quick Start
-
-```bash
-# Download BrowserOS
-browseros-cli install
-
-# Start BrowserOS
-browseros-cli launch
-
-# Auto-configure MCP settings for your AI tools
-browseros-cli init --auto
-
-# Verify everything is working
-browseros-cli health
-```
-
-## Usage
-
-### Navigation
-
-```bash
-browseros-cli navigate "https://example.com"
-```
-
-### Observation
-
-```bash
-browseros-cli snapshot           # Get the accessibility tree of the current page
-browseros-cli console-logs       # View browser console output
-```
-
-### Screenshots
-
-```bash
-browseros-cli screenshot         # Capture the current page
-```
-
-### Input
-
-```bash
-browseros-cli click 42           # Click an element by its node ID
-browseros-cli fill 85 "query"    # Type text into an input field
-```
-
-### Agent Mode
-
-```bash
-browseros-cli agent "Search for flights to Tokyo"
-```
-
-## Documentation
-
-Full documentation is available at [browseros.com](https://browseros.com).
-
-## License
-
-MIT

packages/browseros-agent/apps/cli/npm/bin/browseros-cli.js

@@ -1,32 +0,0 @@
-#!/usr/bin/env node
-
-const { execFileSync, spawnSync } = require('node:child_process')
-const path = require('node:path')
-const fs = require('node:fs')
-
-const BINARY_DIR = path.join(__dirname, '..', '.binary')
-const EXT = process.platform === 'win32' ? '.exe' : ''
-const BIN_PATH = path.join(BINARY_DIR, `browseros-cli${EXT}`)
-
-if (!fs.existsSync(BIN_PATH)) {
-  console.error('browseros-cli: binary not found, downloading...')
-  try {
-    execFileSync(
-      process.execPath,
-      [path.join(__dirname, '..', 'scripts', 'postinstall.js')],
-      { stdio: 'inherit', env: { ...process.env, BROWSEROS_NPM_FORCE: '1' } },
-    )
-  } catch {
-    console.error(
-      'browseros-cli: failed to download binary. Try reinstalling:\n  npm install -g browseros-cli',
-    )
-    process.exit(1)
-  }
-}
-
-const result = spawnSync(BIN_PATH, process.argv.slice(2), {
-  stdio: 'inherit',
-  env: { ...process.env, BROWSEROS_INSTALL_METHOD: 'npm' },
-})
-
-process.exit(result.status ?? 1)

packages/browseros-agent/apps/cli/npm/package.json

@@ -1,45 +0,0 @@
-{
-  "name": "browseros-cli",
-  "version": "0.2.0",
-  "description": "Command-line interface for controlling BrowserOS — launch and automate the browser from the terminal",
-  "bin": {
-    "browseros-cli": "bin/browseros-cli.js"
-  },
-  "scripts": {
-    "postinstall": "node scripts/postinstall.js"
-  },
-  "keywords": [
-    "browseros",
-    "cli",
-    "browser",
-    "automation",
-    "mcp",
-    "ai-agent",
-    "model-context-protocol"
-  ],
-  "repository": {
-    "type": "git",
-    "url": "https://github.com/browseros-ai/BrowserOS",
-    "directory": "packages/browseros-agent/apps/cli/npm"
-  },
-  "homepage": "https://browseros.com",
-  "bugs": "https://github.com/browseros-ai/BrowserOS/issues",
-  "license": "MIT",
-  "os": [
-    "darwin",
-    "linux",
-    "win32"
-  ],
-  "cpu": [
-    "x64",
-    "arm64"
-  ],
-  "engines": {
-    "node": ">=18"
-  },
-  "files": [
-    "bin/",
-    "scripts/",
-    "README.md"
-  ]
-}

packages/browseros-agent/apps/cli/npm/scripts/postinstall.js

@@ -1,142 +0,0 @@
-const https = require('node:https')
-const http = require('node:http')
-const fs = require('node:fs')
-const path = require('node:path')
-const { execSync } = require('node:child_process')
-const { createHash } = require('node:crypto')
-
-const VERSION = require('../package.json').version
-const GITHUB_RELEASE_BASE = `https://github.com/browseros-ai/BrowserOS/releases/download/browseros-cli-v${VERSION}`
-const BINARY_DIR = path.join(__dirname, '..', '.binary')
-const EXT = process.platform === 'win32' ? '.exe' : ''
-const BINARY_PATH = path.join(BINARY_DIR, `browseros-cli${EXT}`)
-
-if (process.env.CI && !process.env.BROWSEROS_NPM_FORCE) {
-  process.exit(0)
-}
-
-const PLATFORM_MAP = { darwin: 'darwin', linux: 'linux', win32: 'windows' }
-const ARCH_MAP = { x64: 'amd64', arm64: 'arm64' }
-
-const platform = PLATFORM_MAP[process.platform]
-const arch = ARCH_MAP[process.arch]
-
-if (!platform || !arch) {
-  console.error(
-    `browseros-cli: unsupported platform ${process.platform}/${process.arch}`,
-  )
-  process.exit(1)
-}
-
-const isWindows = platform === 'windows'
-const archiveExt = isWindows ? 'zip' : 'tar.gz'
-const archiveName = `browseros-cli_${VERSION}_${platform}_${arch}.${archiveExt}`
-const archiveURL = `${GITHUB_RELEASE_BASE}/${archiveName}`
-const checksumURL = `${GITHUB_RELEASE_BASE}/checksums.txt`
-
-const MAX_REDIRECTS = 5
-
-function download(url, redirects = 0) {
-  return new Promise((resolve, reject) => {
-    if (redirects > MAX_REDIRECTS) {
-      return reject(new Error(`Too many redirects for ${url}`))
-    }
-    const client = url.startsWith('https') ? https : http
-    client
-      .get(url, { headers: { 'User-Agent': 'browseros-cli-npm' } }, (res) => {
-        if (
-          res.statusCode >= 300 &&
-          res.statusCode < 400 &&
-          res.headers.location
-        ) {
-          return download(res.headers.location, redirects + 1).then(
-            resolve,
-            reject,
-          )
-        }
-        if (res.statusCode !== 200) {
-          return reject(new Error(`HTTP ${res.statusCode} for ${url}`))
-        }
-        const chunks = []
-        res.on('data', (chunk) => chunks.push(chunk))
-        res.on('end', () => resolve(Buffer.concat(chunks)))
-        res.on('error', reject)
-      })
-      .on('error', reject)
-  })
-}
-
-async function main() {
-  console.log(
-    `browseros-cli: downloading v${VERSION} for ${platform}/${arch}...`,
-  )
-
-  const [archiveBuffer, checksumBuffer] = await Promise.all([
-    download(archiveURL),
-    download(checksumURL).catch(() => null),
-  ])
-
-  if (checksumBuffer) {
-    const checksumText = checksumBuffer.toString('utf-8')
-    const expectedLine = checksumText
-      .split('\n')
-      .find((l) => l.includes(archiveName))
-    if (expectedLine) {
-      const expected = expectedLine.split(/\s+/)[0]
-      const actual = createHash('sha256').update(archiveBuffer).digest('hex')
-      if (actual !== expected) {
-        console.error(
-          `browseros-cli: checksum mismatch!\n  expected: ${expected}\n  got:      ${actual}`,
-        )
-        process.exit(1)
-      }
-      console.log('browseros-cli: checksum verified.')
-    } else {
-      console.warn(
-        'browseros-cli: warning: checksum entry not found in checksums.txt, skipping verification.',
-      )
-    }
-  } else {
-    console.warn(
-      'browseros-cli: warning: could not fetch checksums.txt, skipping verification.',
-    )
-  }
-
-  fs.mkdirSync(BINARY_DIR, { recursive: true })
-  const tmpArchive = path.join(BINARY_DIR, archiveName)
-  fs.writeFileSync(tmpArchive, archiveBuffer)
-
-  if (isWindows) {
-    execSync(
-      `powershell -Command "Expand-Archive -Force -Path '${tmpArchive}' -DestinationPath '${BINARY_DIR}'"`,
-      { stdio: 'inherit' },
-    )
-  } else {
-    execSync(`tar -xzf "${tmpArchive}" -C "${BINARY_DIR}"`, {
-      stdio: 'inherit',
-    })
-  }
-
-  fs.unlinkSync(tmpArchive)
-
-  if (!fs.existsSync(BINARY_PATH)) {
-    console.error(
-      `browseros-cli: binary not found after extraction at ${BINARY_PATH}`,
-    )
-    process.exit(1)
-  }
-
-  if (!isWindows) {
-    fs.chmodSync(BINARY_PATH, 0o755)
-  }
-
-  console.log(`browseros-cli: installed v${VERSION} successfully.`)
-}
-
-main().catch((err) => {
-  console.error(`browseros-cli: installation failed: ${err.message}`)
-  console.error(
-    'You can install manually: curl -fsSL https://cdn.browseros.com/cli/install.sh | bash',
-  )
-  process.exit(1)
-})

packages/browseros-agent/apps/cli/scripts/install.ps1

@@ -1,115 +0,0 @@
-#
-# Install browseros-cli for Windows — downloads the latest release binary.
-#
-# Usage (PowerShell — save and run):
-#   Invoke-WebRequest -Uri "https://cdn.browseros.com/cli/install.ps1" -OutFile install.ps1
-#   .\install.ps1
-#   .\install.ps1 -Version "0.1.0" -Dir "C:\tools\browseros"
-#
-# Usage (one-liner, uses env vars for options):
-#   & { $env:BROWSEROS_VERSION="0.1.0"; irm https://cdn.browseros.com/cli/install.ps1 | iex }
-#
-
-param(
-    [string]$Version = "",
-    [string]$Dir = ""
-)
-
-$ErrorActionPreference = "Stop"
-
-# TLS 1.2 — older PS 5.1 defaults to TLS 1.0
-[Net.ServicePointManager]::SecurityProtocol = [Net.SecurityProtocolType]::Tls12
-
-$CdnBase = "https://cdn.browseros.com/cli"
-$Binary = "browseros-cli"
-
-# When piped via irm | iex, param() is ignored — fall back to env vars
-if (-not $Version) { $Version = $env:BROWSEROS_VERSION }
-if (-not $Dir) { $Dir = if ($env:BROWSEROS_DIR) { $env:BROWSEROS_DIR } else { "$env:LOCALAPPDATA\browseros-cli\bin" } }
-
-# ── Resolve latest version ───────────────────────────────────────────────────
-
-if (-not $Version) {
-    Write-Host "Fetching latest version..."
-    $Version = (Invoke-WebRequest -Uri "$CdnBase/latest/version.txt" -UseBasicParsing).Content.Trim()
-    if (-not $Version) {
-        Write-Error "Could not determine latest version. Try: -Version 0.1.0"
-        exit 1
-    }
-}
-
-if ($Version -notmatch '^\d+\.\d+\.\d+(-[a-zA-Z0-9.]+)?$') {
-    Write-Error "Unexpected version format: '$Version'"
-    exit 1
-}
-
-Write-Host "Installing browseros-cli v$Version..."
-
-# ── Detect architecture ──────────────────────────────────────────────────────
-
-# $env:PROCESSOR_ARCHITECTURE lies under x64 emulation on ARM64 Windows.
-# Use .NET RuntimeInformation when available, fall back to PROCESSOR_ARCHITEW6432.
-$Arch = "amd64"
-try {
-    $osArch = [System.Runtime.InteropServices.RuntimeInformation]::OSArchitecture
-    if ($osArch -eq [System.Runtime.InteropServices.Architecture]::Arm64) { $Arch = "arm64" }
-} catch {
-    if ($env:PROCESSOR_ARCHITEW6432 -eq "ARM64" -or $env:PROCESSOR_ARCHITECTURE -eq "ARM64") {
-        $Arch = "arm64"
-    }
-}
-
-if (-not [Environment]::Is64BitOperatingSystem) {
-    Write-Error "32-bit Windows is not supported."
-    exit 1
-}
-
-# ── Download and extract ─────────────────────────────────────────────────────
-
-$Filename = "${Binary}_${Version}_windows_${Arch}.zip"
-$Url = "$CdnBase/v$Version/$Filename"
-$TmpDir = Join-Path ([System.IO.Path]::GetTempPath()) ("browseros-cli-install-" + [System.IO.Path]::GetRandomFileName())
-
-try {
-    New-Item -ItemType Directory -Path $TmpDir | Out-Null
-
-    $ZipPath = Join-Path $TmpDir $Filename
-
-    Write-Host "Downloading $Url..."
-    Invoke-WebRequest -Uri $Url -OutFile $ZipPath -UseBasicParsing
-
-    Expand-Archive -Path $ZipPath -DestinationPath $TmpDir -Force
-
-    $Exe = Get-ChildItem -Path $TmpDir -Filter "$Binary.exe" -File -Recurse | Select-Object -First 1
-    if (-not $Exe) {
-        Write-Error "Binary not found in archive."
-        exit 1
-    }
-
-    # ── Install ──────────────────────────────────────────────────────────────
-
-    if (-not (Test-Path $Dir)) {
-        New-Item -ItemType Directory -Path $Dir -Force | Out-Null
-    }
-
-    Move-Item -Force $Exe.FullName (Join-Path $Dir "$Binary.exe")
-
-    Write-Host "Installed $Binary.exe to $Dir"
-} finally {
-    if (Test-Path $TmpDir) { Remove-Item -Recurse -Force $TmpDir -ErrorAction SilentlyContinue }
-}
-
-# ── PATH ─────────────────────────────────────────────────────────────────────
-
-$UserPath = [Environment]::GetEnvironmentVariable("Path", "User")
-$PathEntries = $UserPath -split ";" | Where-Object { $_ -ne "" }
-if ($Dir -notin $PathEntries) {
-    Write-Host ""
-    Write-Host "Adding $Dir to your user PATH..."
-    [Environment]::SetEnvironmentVariable("Path", "$Dir;$UserPath", "User")
-    $env:Path = "$Dir;$env:Path"
-    Write-Host "Done. Restart your terminal for PATH changes to take effect."
-}
-
-Write-Host ""
-Write-Host "Run 'browseros-cli --help' to get started."

packages/browseros-agent/apps/cli/scripts/install.sh

@@ -1,151 +0,0 @@
-#!/usr/bin/env bash
-#
-# Install browseros-cli — downloads the latest release binary for your platform.
-#
-# Usage:
-#   curl -fsSL https://cdn.browseros.com/cli/install.sh | bash
-#
-#   # Or with options:
-#   curl -fsSL https://cdn.browseros.com/cli/install.sh | bash -s -- --version 0.1.0 --dir /usr/local/bin
-
-set -euo pipefail
-
-CDN_BASE="https://cdn.browseros.com/cli"
-BINARY="browseros-cli"
-INSTALL_DIR="${HOME}/.browseros/bin"
-
-# ── Parse arguments ──────────────────────────────────────────────────────────
-
-VERSION=""
-CUSTOM_DIR=""
-
-while [[ $# -gt 0 ]]; do
-  case "$1" in
-    --version)
-      [[ $# -lt 2 ]] && { echo "Error: --version requires a value" >&2; exit 1; }
-      VERSION="$2"; shift 2 ;;
-    --dir)
-      [[ $# -lt 2 ]] && { echo "Error: --dir requires a value" >&2; exit 1; }
-      CUSTOM_DIR="$2"; shift 2 ;;
-    --help)
-      echo "Usage: install.sh [--version VERSION] [--dir INSTALL_DIR]"
-      echo ""
-      echo "  --version   Install a specific version (default: latest)"
-      echo "  --dir       Install directory (default: ~/.browseros/bin)"
-      exit 0
-      ;;
-    *) echo "Unknown option: $1" >&2; exit 1 ;;
-  esac
-done
-
-[[ -n "$CUSTOM_DIR" ]] && INSTALL_DIR="$CUSTOM_DIR"
-
-# ── Resolve latest version ───────────────────────────────────────────────────
-
-if [[ -z "$VERSION" ]]; then
-  VERSION=$(curl -fsSL "${CDN_BASE}/latest/version.txt" | tr -d '[:space:]')
-
-  if [[ -z "$VERSION" ]]; then
-    echo "Error: could not determine latest version." >&2
-    echo "  Try: install.sh --version 0.1.0" >&2
-    exit 1
-  fi
-fi
-
-if [[ ! "$VERSION" =~ ^[0-9]+\.[0-9]+\.[0-9]+(-[a-zA-Z0-9.]+)?$ ]]; then
-  echo "Error: unexpected version format: '$VERSION'" >&2
-  exit 1
-fi
-
-echo "Installing browseros-cli v${VERSION}..."
-
-# ── Detect platform ──────────────────────────────────────────────────────────
-
-OS=$(uname -s | tr '[:upper:]' '[:lower:]')
-ARCH=$(uname -m)
-
-case "$OS" in
-  darwin) OS="darwin" ;;
-  linux)  OS="linux" ;;
-  *)      echo "Error: unsupported OS: $OS" >&2; exit 1 ;;
-esac
-
-case "$ARCH" in
-  x86_64|amd64) ARCH="amd64" ;;
-  arm64|aarch64) ARCH="arm64" ;;
-  *)             echo "Error: unsupported architecture: $ARCH" >&2; exit 1 ;;
-esac
-
-# ── Download and extract ─────────────────────────────────────────────────────
-
-FILENAME="${BINARY}_${VERSION}_${OS}_${ARCH}.tar.gz"
-URL="${CDN_BASE}/v${VERSION}/${FILENAME}"
-CHECKSUM_URL="${CDN_BASE}/v${VERSION}/checksums.txt"
-
-TMPDIR_DL=$(mktemp -d)
-trap 'rm -rf "$TMPDIR_DL"' EXIT
-
-echo "Downloading ${URL}..."
-curl -fSL --progress-bar -o "${TMPDIR_DL}/${FILENAME}" "$URL"
-
-# Verify checksum if sha256sum/shasum is available
-if curl -fsSL -o "${TMPDIR_DL}/checksums.txt" "$CHECKSUM_URL" 2>/dev/null; then
-  expected=$(awk -v filename="$FILENAME" '$2 == filename { print $1; exit }' "${TMPDIR_DL}/checksums.txt")
-  if [[ -n "$expected" ]]; then
-    if command -v sha256sum >/dev/null 2>&1; then
-      actual=$(sha256sum "${TMPDIR_DL}/${FILENAME}" | awk '{print $1}')
-    elif command -v shasum >/dev/null 2>&1; then
-      actual=$(shasum -a 256 "${TMPDIR_DL}/${FILENAME}" | awk '{print $1}')
-    else
-      actual=""
-      echo "Warning: no sha256sum/shasum found; skipping checksum verification." >&2
-    fi
-    if [[ -n "$actual" && "$actual" != "$expected" ]]; then
-      echo "Error: checksum mismatch (expected ${expected}, got ${actual})" >&2
-      exit 1
-    fi
-    [[ -n "$actual" ]] && echo "Checksum verified."
-  else
-    echo "Warning: checksum not found in checksums.txt; skipping verification." >&2
-  fi
-else
-  echo "Warning: could not fetch checksums.txt; skipping checksum verification." >&2
-fi
-
-tar -xzf "${TMPDIR_DL}/${FILENAME}" -C "$TMPDIR_DL"
-
-BINARY_PATH="${TMPDIR_DL}/${BINARY}"
-if [[ ! -f "$BINARY_PATH" ]]; then
-  BINARY_PATH=$(find "$TMPDIR_DL" -type f -name "$BINARY" -print -quit)
-fi
-
-if [[ -z "$BINARY_PATH" || ! -f "$BINARY_PATH" ]]; then
-  echo "Error: binary not found in archive." >&2
-  exit 1
-fi
-
-# ── Install ──────────────────────────────────────────────────────────────────
-
-mkdir -p "$INSTALL_DIR"
-mv "$BINARY_PATH" "${INSTALL_DIR}/${BINARY}"
-chmod +x "${INSTALL_DIR}/${BINARY}"
-
-echo "Installed ${BINARY} to ${INSTALL_DIR}/${BINARY}"
-
-# ── PATH hint ────────────────────────────────────────────────────────────────
-
-if ! echo "$PATH" | tr ':' '\n' | grep -qx "$INSTALL_DIR"; then
-  echo ""
-  echo "Add browseros-cli to your PATH:"
-  echo ""
-
-  SHELL_NAME=$(basename "${SHELL:-/bin/bash}")
-  case "$SHELL_NAME" in
-    zsh)  echo "  echo 'export PATH=\"${INSTALL_DIR}:\$PATH\"' >> ~/.zshrc && source ~/.zshrc" ;;
-    fish) echo "  fish_add_path ${INSTALL_DIR}" ;;
-    *)    echo "  echo 'export PATH=\"${INSTALL_DIR}:\$PATH\"' >> ~/.bashrc && source ~/.bashrc" ;;
-  esac
-fi
-
-echo ""
-echo "Run 'browseros-cli --help' to get started."

packages/browseros-agent/apps/cli/update/apply.go

@@ -1,49 +0,0 @@
-package update
-
-import (
-	"bytes"
-	"crypto/sha256"
-	"encoding/hex"
-	"fmt"
-	"strings"
-
-	"github.com/minio/selfupdate"
-)
-
-func CheckPermissions(targetPath string) error {
-	options := selfupdate.Options{TargetPath: targetPath}
-	return options.CheckPermissions()
-}
-
-func VerifyChecksum(data []byte, expectedHex string) error {
-	expected, err := decodeChecksum(expectedHex)
-	if err != nil {
-		return err
-	}
-	actual := sha256.Sum256(data)
-	if !bytes.Equal(actual[:], expected) {
-		return fmt.Errorf(
-			"checksum mismatch: expected %s, got %s",
-			hex.EncodeToString(expected),
-			hex.EncodeToString(actual[:]),
-		)
-	}
-	return nil
-}
-
-func ApplyBinary(binary []byte, targetPath string) error {
-	options := selfupdate.Options{TargetPath: targetPath}
-	err := selfupdate.Apply(bytes.NewReader(binary), options)
-	if rollbackErr := selfupdate.RollbackError(err); rollbackErr != nil {
-		return fmt.Errorf("update failed and rollback failed: %w", rollbackErr)
-	}
-	return err
-}
-
-func decodeChecksum(checksumHex string) ([]byte, error) {
-	value := strings.TrimSpace(checksumHex)
-	if value == "" {
-		return nil, fmt.Errorf("missing checksum")
-	}
-	return hex.DecodeString(value)
-}

packages/browseros-agent/apps/cli/update/archive.go

@@ -1,138 +0,0 @@
-package update
-
-import (
-	"archive/tar"
-	"archive/zip"
-	"bytes"
-	"compress/gzip"
-	"context"
-	"fmt"
-	"io"
-	"net/http"
-)
-
-const maxAssetSize = 64 << 20
-const maxBinarySize = 256 << 20
-
-func DownloadAsset(ctx context.Context, client *http.Client, asset Asset) ([]byte, error) {
-	req, err := http.NewRequestWithContext(ctx, http.MethodGet, asset.URL, nil)
-	if err != nil {
-		return nil, err
-	}
-
-	resp, err := client.Do(req)
-	if err != nil {
-		return nil, err
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("update download returned HTTP %d", resp.StatusCode)
-	}
-
-	return readAssetBytes(resp.Body)
-}
-
-func readAssetBytes(reader io.Reader) ([]byte, error) {
-	limited := io.LimitReader(reader, maxAssetSize+1)
-	data, err := io.ReadAll(limited)
-	if err != nil {
-		return nil, err
-	}
-	if len(data) > maxAssetSize {
-		return nil, fmt.Errorf("update asset exceeds %d bytes", maxAssetSize)
-	}
-	return data, nil
-}
-
-func ExtractBinary(archive []byte, format string) ([]byte, error) {
-	switch format {
-	case "tar.gz":
-		return extractTarGzBinary(archive)
-	case "zip":
-		return extractZipBinary(archive)
-	default:
-		return nil, fmt.Errorf("unsupported archive format %q", format)
-	}
-}
-
-func extractTarGzBinary(archive []byte) ([]byte, error) {
-	gzipReader, err := gzip.NewReader(bytes.NewReader(archive))
-	if err != nil {
-		return nil, err
-	}
-	defer gzipReader.Close()
-
-	tarReader := tar.NewReader(gzipReader)
-	return readTarBinary(tarReader)
-}
-
-func readTarBinary(reader *tar.Reader) ([]byte, error) {
-	var binary []byte
-
-	for {
-		header, err := reader.Next()
-		if err == io.EOF {
-			break
-		}
-		if err != nil {
-			return nil, err
-		}
-		if header.Typeflag != tar.TypeReg {
-			continue
-		}
-		if binary != nil {
-			return nil, fmt.Errorf("archive contains multiple files; expected exactly one binary")
-		}
-
-		binary, err = io.ReadAll(io.LimitReader(reader, maxBinarySize+1))
-		if err != nil {
-			return nil, err
-		}
-		if len(binary) > maxBinarySize {
-			return nil, fmt.Errorf("extracted binary exceeds %d bytes", maxBinarySize)
-		}
-	}
-
-	if binary == nil {
-		return nil, fmt.Errorf("archive does not contain a file")
-	}
-
-	return binary, nil
-}
-
-func extractZipBinary(archive []byte) ([]byte, error) {
-	reader, err := zip.NewReader(bytes.NewReader(archive), int64(len(archive)))
-	if err != nil {
-		return nil, err
-	}
-
-	var binary []byte
-	for _, file := range reader.File {
-		if file.FileInfo().IsDir() {
-			continue
-		}
-		if binary != nil {
-			return nil, fmt.Errorf("archive contains multiple files; expected exactly one binary")
-		}
-
-		rc, err := file.Open()
-		if err != nil {
-			return nil, err
-		}
-		binary, err = io.ReadAll(io.LimitReader(rc, maxBinarySize+1))
-		rc.Close()
-		if err != nil {
-			return nil, err
-		}
-		if len(binary) > maxBinarySize {
-			return nil, fmt.Errorf("extracted binary exceeds %d bytes", maxBinarySize)
-		}
-	}
-
-	if binary == nil {
-		return nil, fmt.Errorf("archive does not contain a file")
-	}
-
-	return binary, nil
-}

packages/browseros-agent/apps/cli/update/archive_test.go

@@ -1,168 +0,0 @@
-package update
-
-import (
-	"archive/tar"
-	"archive/zip"
-	"bytes"
-	"compress/gzip"
-	"crypto/sha256"
-	"encoding/hex"
-	"os"
-	"path/filepath"
-	"testing"
-)
-
-func TestExtractBinaryTarGz(t *testing.T) {
-	archive := createTarGz(t, map[string]string{"browseros-cli": "new-binary"})
-
-	binary, err := ExtractBinary(archive, "tar.gz")
-	if err != nil {
-		t.Fatalf("ExtractBinary() error = %v", err)
-	}
-	if string(binary) != "new-binary" {
-		t.Fatalf("ExtractBinary() = %q, want %q", string(binary), "new-binary")
-	}
-}
-
-func TestExtractBinaryZip(t *testing.T) {
-	archive := createZip(t, map[string]string{"browseros-cli.exe": "new-binary"})
-
-	binary, err := ExtractBinary(archive, "zip")
-	if err != nil {
-		t.Fatalf("ExtractBinary() error = %v", err)
-	}
-	if string(binary) != "new-binary" {
-		t.Fatalf("ExtractBinary() = %q, want %q", string(binary), "new-binary")
-	}
-}
-
-func TestExtractBinaryTarGzRejectsMultipleFiles(t *testing.T) {
-	archive := createTarGz(t, map[string]string{
-		"browseros-cli":     "new-binary",
-		"browseros-cli.sig": "signature",
-	})
-
-	_, err := ExtractBinary(archive, "tar.gz")
-	if err == nil {
-		t.Fatal("ExtractBinary() error = nil, want multiple files error")
-	}
-	if err.Error() != "archive contains multiple files; expected exactly one binary" {
-		t.Fatalf("ExtractBinary() error = %q", err)
-	}
-}
-
-func TestVerifyChecksumValid(t *testing.T) {
-	data := []byte("some-data")
-	sum := sha256.Sum256(data)
-	if err := VerifyChecksum(data, hex.EncodeToString(sum[:])); err != nil {
-		t.Fatalf("VerifyChecksum() error = %v", err)
-	}
-}
-
-func TestVerifyChecksumMismatch(t *testing.T) {
-	data := []byte("some-data")
-	badChecksum := "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
-	if err := VerifyChecksum(data, badChecksum); err == nil {
-		t.Fatal("VerifyChecksum() error = nil, want mismatch error")
-	}
-}
-
-func TestApplyBinary(t *testing.T) {
-	targetPath := filepath.Join(t.TempDir(), "browseros-cli")
-	if err := os.WriteFile(targetPath, []byte("old-binary"), 0755); err != nil {
-		t.Fatalf("WriteFile() error = %v", err)
-	}
-
-	newBinary := []byte("new-binary")
-	if err := ApplyBinary(newBinary, targetPath); err != nil {
-		t.Fatalf("ApplyBinary() error = %v", err)
-	}
-
-	data, err := os.ReadFile(targetPath)
-	if err != nil {
-		t.Fatalf("ReadFile() error = %v", err)
-	}
-	if string(data) != "new-binary" {
-		t.Fatalf("updated binary = %q, want %q", string(data), "new-binary")
-	}
-}
-
-func TestVerifyThenApplyIntegration(t *testing.T) {
-	archive := createTarGz(t, map[string]string{"browseros-cli": "updated-binary"})
-	archiveSum := sha256.Sum256(archive)
-
-	if err := VerifyChecksum(archive, hex.EncodeToString(archiveSum[:])); err != nil {
-		t.Fatalf("VerifyChecksum(archive) error = %v", err)
-	}
-
-	binary, err := ExtractBinary(archive, "tar.gz")
-	if err != nil {
-		t.Fatalf("ExtractBinary() error = %v", err)
-	}
-
-	targetPath := filepath.Join(t.TempDir(), "browseros-cli")
-	if err := os.WriteFile(targetPath, []byte("old"), 0755); err != nil {
-		t.Fatalf("WriteFile() error = %v", err)
-	}
-	if err := ApplyBinary(binary, targetPath); err != nil {
-		t.Fatalf("ApplyBinary() error = %v", err)
-	}
-
-	data, err := os.ReadFile(targetPath)
-	if err != nil {
-		t.Fatalf("ReadFile() error = %v", err)
-	}
-	if string(data) != "updated-binary" {
-		t.Fatalf("binary = %q, want %q", string(data), "updated-binary")
-	}
-}
-
-func createTarGz(t *testing.T, files map[string]string) []byte {
-	t.Helper()
-
-	var buffer bytes.Buffer
-	gzipWriter := gzip.NewWriter(&buffer)
-	tarWriter := tar.NewWriter(gzipWriter)
-	for name, body := range files {
-		data := []byte(body)
-		if err := tarWriter.WriteHeader(&tar.Header{
-			Name: name,
-			Mode: 0755,
-			Size: int64(len(data)),
-		}); err != nil {
-			t.Fatalf("WriteHeader() error = %v", err)
-		}
-		if _, err := tarWriter.Write(data); err != nil {
-			t.Fatalf("Write() error = %v", err)
-		}
-	}
-	if err := tarWriter.Close(); err != nil {
-		t.Fatalf("Close() error = %v", err)
-	}
-	if err := gzipWriter.Close(); err != nil {
-		t.Fatalf("Close() error = %v", err)
-	}
-
-	return buffer.Bytes()
-}
-
-func createZip(t *testing.T, files map[string]string) []byte {
-	t.Helper()
-
-	var buffer bytes.Buffer
-	zipWriter := zip.NewWriter(&buffer)
-	for name, body := range files {
-		fileWriter, err := zipWriter.Create(name)
-		if err != nil {
-			t.Fatalf("Create() error = %v", err)
-		}
-		if _, err := fileWriter.Write([]byte(body)); err != nil {
-			t.Fatalf("Write() error = %v", err)
-		}
-	}
-	if err := zipWriter.Close(); err != nil {
-		t.Fatalf("Close() error = %v", err)
-	}
-
-	return buffer.Bytes()
-}

packages/browseros-agent/apps/cli/update/manager.go

@@ -1,273 +0,0 @@
-package update
-
-import (
-	"context"
-	"fmt"
-	"net/http"
-	"os"
-	"runtime"
-	"time"
-)
-
-const (
-	DefaultManifestURL     = "https://cdn.browseros.com/cli/latest/manifest.json"
-	DefaultCheckTTL        = 24 * time.Hour
-	DefaultHTTPTimeout     = 2 * time.Second
-	DefaultDownloadTimeout = 5 * time.Minute
-	SkipCheckEnv           = "BROWSEROS_SKIP_UPDATE_CHECK"
-	InstallMethodEnv       = "BROWSEROS_INSTALL_METHOD"
-)
-
-type Options struct {
-	CurrentVersion  string
-	ManifestURL     string
-	CheckTTL        time.Duration
-	HTTPTimeout     time.Duration
-	DownloadTimeout time.Duration
-	JSONOutput      bool
-	Debug           bool
-	Automatic       bool
-	HTTPClient      *http.Client
-	Now             func() time.Time
-}
-
-type Manager struct {
-	options Options
-	state   *State
-}
-
-type CheckResult struct {
-	CurrentVersion    string    `json:"current_version"`
-	LatestVersion     string    `json:"latest_version"`
-	LatestPublishedAt string    `json:"latest_published_at,omitempty"`
-	UpdateAvailable   bool      `json:"update_available"`
-	CheckedAt         time.Time `json:"checked_at"`
-	Asset             *Asset    `json:"asset,omitempty"`
-}
-
-func NewManager(options Options) *Manager {
-	if options.ManifestURL == "" {
-		options.ManifestURL = DefaultManifestURL
-	}
-	if options.CheckTTL == 0 {
-		options.CheckTTL = DefaultCheckTTL
-	}
-	if options.HTTPTimeout == 0 {
-		options.HTTPTimeout = DefaultHTTPTimeout
-	}
-	if options.DownloadTimeout == 0 {
-		options.DownloadTimeout = DefaultDownloadTimeout
-	}
-	if options.Now == nil {
-		options.Now = time.Now
-	}
-	if options.HTTPClient == nil {
-		options.HTTPClient = &http.Client{}
-	}
-
-	state, err := LoadState()
-	if err != nil {
-		state = &State{}
-	}
-
-	return &Manager{
-		options: options,
-		state:   state,
-	}
-}
-
-func (m *Manager) CachedNotice() string {
-	if !m.AutomaticEnabled() || m.state == nil || m.state.LatestVersion == "" {
-		return ""
-	}
-
-	comparison, err := CompareVersions(m.options.CurrentVersion, m.state.LatestVersion)
-	if err != nil || comparison >= 0 {
-		return ""
-	}
-
-	return FormatNotice(m.options.CurrentVersion, m.state.LatestVersion)
-}
-
-func (m *Manager) AutomaticEnabled() bool {
-	if !m.options.Automatic || m.options.JSONOutput {
-		return false
-	}
-	if os.Getenv(SkipCheckEnv) != "" {
-		return false
-	}
-	if installedViaPackageManager() {
-		return false
-	}
-	return IsReleaseVersion(m.options.CurrentVersion)
-}
-
-func installedViaPackageManager() bool {
-	method := os.Getenv(InstallMethodEnv)
-	switch method {
-	case "npm", "brew", "homebrew":
-		return true
-	}
-	return false
-}
-
-func (m *Manager) ShouldCheck() bool {
-	if !m.AutomaticEnabled() {
-		return false
-	}
-	return m.state.IsStale(m.options.Now(), m.options.CheckTTL)
-}
-
-func (m *Manager) StartBackgroundCheck(ctx context.Context) <-chan struct{} {
-	done := make(chan struct{})
-	if !m.ShouldCheck() {
-		close(done)
-		return done
-	}
-
-	go func() {
-		defer close(done)
-		_, _ = m.CheckNow(ctx)
-	}()
-
-	return done
-}
-
-func (m *Manager) CheckNow(ctx context.Context) (*CheckResult, error) {
-	if !IsReleaseVersion(m.options.CurrentVersion) {
-		return nil, fmt.Errorf("self-update is unavailable for non-release build %q", m.options.CurrentVersion)
-	}
-
-	checkCtx, cancel := context.WithTimeout(ctx, m.options.HTTPTimeout)
-	defer cancel()
-
-	manifest, err := FetchManifest(checkCtx, cloneHTTPClient(m.options.HTTPClient), m.options.ManifestURL)
-	if err != nil {
-		m.recordError(err)
-		return nil, err
-	}
-
-	asset, err := SelectAsset(manifest, runtime.GOOS, runtime.GOARCH)
-	if err != nil {
-		m.recordError(err)
-		return nil, err
-	}
-
-	comparison, err := CompareVersions(m.options.CurrentVersion, manifest.Version)
-	if err != nil {
-		m.recordError(err)
-		return nil, err
-	}
-
-	result := &CheckResult{
-		CurrentVersion:    m.options.CurrentVersion,
-		LatestVersion:     manifest.Version,
-		LatestPublishedAt: manifest.PublishedAt,
-		UpdateAvailable:   comparison < 0,
-		CheckedAt:         m.options.Now(),
-	}
-	if result.UpdateAvailable {
-		assetCopy := asset
-		result.Asset = &assetCopy
-	}
-
-	m.state = &State{
-		LastCheckedAt:     result.CheckedAt,
-		LatestVersion:     manifest.Version,
-		LatestPublishedAt: manifest.PublishedAt,
-		AssetURL:          asset.URL,
-	}
-	_ = SaveState(m.state)
-
-	return result, nil
-}
-
-func (m *Manager) Apply(ctx context.Context, result *CheckResult) error {
-	if result == nil || !result.UpdateAvailable || result.Asset == nil {
-		return fmt.Errorf("browseros-cli is already up to date")
-	}
-
-	downloadCtx, cancel := context.WithTimeout(ctx, m.options.DownloadTimeout)
-	defer cancel()
-
-	archive, err := DownloadAsset(downloadCtx, cloneHTTPClient(m.options.HTTPClient), *result.Asset)
-	if err != nil {
-		return err
-	}
-	if err := VerifyChecksum(archive, result.Asset.SHA256); err != nil {
-		return err
-	}
-	binary, err := ExtractBinary(archive, result.Asset.ArchiveFormat)
-	if err != nil {
-		return err
-	}
-
-	targetPath, err := os.Executable()
-	if err != nil {
-		return err
-	}
-	if err := CheckPermissions(targetPath); err != nil {
-		return fmt.Errorf(
-			"cannot replace %s: %w\n\nReinstall with the installer script or move the binary to a writable location.",
-			targetPath,
-			err,
-		)
-	}
-	if err := ApplyBinary(binary, targetPath); err != nil {
-		return err
-	}
-
-	m.saveAppliedState(result)
-
-	return nil
-}
-
-func FormatNotice(currentVersion, latestVersion string) string {
-	notice := fmt.Sprintf(
-		"Update available: browseros-cli v%s (current v%s)",
-		latestVersion,
-		currentVersion,
-	)
-
-	switch os.Getenv(InstallMethodEnv) {
-	case "npm":
-		notice += "\nRun `npm update -g browseros-cli` to upgrade."
-	case "brew", "homebrew":
-		notice += "\nRun `brew upgrade browseros-cli` to upgrade."
-	default:
-		notice += "\nRun `browseros-cli update` to upgrade."
-	}
-
-	return notice
-}
-
-func (m *Manager) recordError(err error) {
-	state := &State{}
-	if m.state != nil {
-		*state = *m.state
-	}
-	state.CheckError = err.Error()
-	m.state = state
-	_ = SaveState(state)
-}
-
-func (m *Manager) saveAppliedState(result *CheckResult) {
-	state := &State{
-		LastCheckedAt:     m.options.Now(),
-		LatestVersion:     result.LatestVersion,
-		LatestPublishedAt: result.LatestPublishedAt,
-		AssetURL:          result.Asset.URL,
-	}
-	m.state = state
-	_ = SaveState(state)
-}
-
-func cloneHTTPClient(client *http.Client) *http.Client {
-	if client == nil {
-		return &http.Client{}
-	}
-
-	cloned := *client
-	cloned.Timeout = 0
-	return &cloned
-}

packages/browseros-agent/apps/cli/update/manager_test.go

@@ -1,188 +0,0 @@
-package update
-
-import (
-	"context"
-	"net/http"
-	"net/http/httptest"
-	"runtime"
-	"testing"
-	"time"
-)
-
-func TestManagerCachedNotice(t *testing.T) {
-	manager := NewManager(Options{
-		CurrentVersion: "1.0.0",
-		Automatic:      true,
-	})
-	manager.state = &State{LatestVersion: "1.2.0"}
-
-	notice := manager.CachedNotice()
-	if notice == "" {
-		t.Fatal("CachedNotice() returned empty notice")
-	}
-}
-
-func TestManagerShouldCheck(t *testing.T) {
-	manager := NewManager(Options{
-		CurrentVersion: "1.0.0",
-		Automatic:      true,
-		CheckTTL:       time.Minute,
-		Now: func() time.Time {
-			return time.Unix(1000, 0).UTC()
-		},
-	})
-	manager.state = &State{LastCheckedAt: time.Unix(0, 0).UTC()}
-
-	if !manager.ShouldCheck() {
-		t.Fatal("ShouldCheck() = false, want true")
-	}
-}
-
-func TestManagerCheckNow(t *testing.T) {
-	configRoot := t.TempDir()
-	t.Setenv("XDG_CONFIG_HOME", configRoot)
-
-	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "application/json")
-		_, _ = w.Write([]byte(`{
-			"version":"9.9.9",
-			"published_at":"2026-03-27T19:00:00Z",
-			"tag":"browseros-cli-v9.9.9",
-			"assets":{
-				"` + runtimePlatformKey(t) + `":{
-					"filename":"browseros-cli_9.9.9_test.tar.gz",
-					"url":"https://cdn.example.com/cli/v9.9.9/browseros-cli_9.9.9_test.tar.gz",
-					"archive_format":"tar.gz",
-					"sha256":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
-				}
-			}
-		}`))
-	}))
-	defer server.Close()
-
-	manager := NewManager(Options{
-		CurrentVersion: "1.0.0",
-		ManifestURL:    server.URL,
-		Automatic:      false,
-		HTTPClient:     server.Client(),
-		Now: func() time.Time {
-			return time.Unix(100, 0).UTC()
-		},
-	})
-
-	result, err := manager.CheckNow(context.Background())
-	if err != nil {
-		t.Fatalf("CheckNow() error = %v", err)
-	}
-	if !result.UpdateAvailable {
-		t.Fatal("CheckNow() UpdateAvailable = false, want true")
-	}
-	if result.LatestPublishedAt != "2026-03-27T19:00:00Z" {
-		t.Fatalf(
-			"CheckNow() LatestPublishedAt = %q, want %q",
-			result.LatestPublishedAt,
-			"2026-03-27T19:00:00Z",
-		)
-	}
-	if manager.state.LatestPublishedAt != "2026-03-27T19:00:00Z" {
-		t.Fatalf(
-			"state LatestPublishedAt = %q, want %q",
-			manager.state.LatestPublishedAt,
-			"2026-03-27T19:00:00Z",
-		)
-	}
-}
-
-func TestCloneHTTPClientClearsTimeout(t *testing.T) {
-	base := &http.Client{Timeout: time.Second}
-
-	cloned := cloneHTTPClient(base)
-
-	if cloned == base {
-		t.Fatal("cloneHTTPClient() returned the original client")
-	}
-	if cloned.Timeout != 0 {
-		t.Fatalf("cloneHTTPClient() Timeout = %s, want 0", cloned.Timeout)
-	}
-	if base.Timeout != time.Second {
-		t.Fatalf("base Timeout = %s, want %s", base.Timeout, time.Second)
-	}
-}
-
-func TestManagerSaveAppliedState(t *testing.T) {
-	configRoot := t.TempDir()
-	t.Setenv("XDG_CONFIG_HOME", configRoot)
-
-	now := time.Unix(200, 0).UTC()
-	manager := NewManager(Options{
-		CurrentVersion: "1.0.0",
-		Now: func() time.Time {
-			return now
-		},
-	})
-	manager.state = &State{
-		LastCheckedAt: time.Unix(100, 0).UTC(),
-		CheckError:    "manifest fetch failed",
-	}
-
-	manager.saveAppliedState(&CheckResult{
-		LatestVersion:     "9.9.9",
-		LatestPublishedAt: "2026-03-27T19:00:00Z",
-		Asset: &Asset{
-			URL: "https://cdn.example.com/cli/v9.9.9/browseros-cli_9.9.9_test.tar.gz",
-		},
-	})
-
-	if manager.state.LastCheckedAt != now {
-		t.Fatalf("LastCheckedAt = %v, want %v", manager.state.LastCheckedAt, now)
-	}
-	if manager.state.CheckError != "" {
-		t.Fatalf("CheckError = %q, want empty", manager.state.CheckError)
-	}
-	if manager.state.LatestPublishedAt != "2026-03-27T19:00:00Z" {
-		t.Fatalf("LatestPublishedAt = %q", manager.state.LatestPublishedAt)
-	}
-}
-
-func TestAutomaticEnabledSkipsForPackageManagerInstall(t *testing.T) {
-	t.Setenv("BROWSEROS_INSTALL_METHOD", "npm")
-
-	manager := NewManager(Options{
-		CurrentVersion: "1.0.0",
-		Automatic:      true,
-	})
-
-	if manager.AutomaticEnabled() {
-		t.Fatal("AutomaticEnabled() = true, want false when BROWSEROS_INSTALL_METHOD=npm")
-	}
-}
-
-func TestAutomaticEnabledAllowsNormalInstall(t *testing.T) {
-	t.Setenv("BROWSEROS_INSTALL_METHOD", "")
-
-	manager := NewManager(Options{
-		CurrentVersion: "1.0.0",
-		Automatic:      true,
-	})
-
-	if !manager.AutomaticEnabled() {
-		t.Fatal("AutomaticEnabled() = false, want true when BROWSEROS_INSTALL_METHOD is empty")
-	}
-}
-
-func runtimePlatformKey(t *testing.T) string {
-	t.Helper()
-	key, err := PlatformKey(runtimeGOOS(), runtimeGOARCH())
-	if err != nil {
-		t.Fatalf("PlatformKey() error = %v", err)
-	}
-	return key
-}
-
-func runtimeGOOS() string {
-	return runtime.GOOS
-}
-
-func runtimeGOARCH() string {
-	return runtime.GOARCH
-}

packages/browseros-agent/apps/cli/update/manifest.go

@@ -1,144 +0,0 @@
-package update
-
-import (
-	"context"
-	"encoding/json"
-	"fmt"
-	"io"
-	"net/http"
-	"strings"
-
-	"golang.org/x/mod/semver"
-)
-
-const maxManifestSize = 1 << 20
-
-type Manifest struct {
-	Version     string           `json:"version"`
-	PublishedAt string           `json:"published_at"`
-	Tag         string           `json:"tag"`
-	Assets      map[string]Asset `json:"assets"`
-}
-
-type Asset struct {
-	Filename      string `json:"filename"`
-	URL           string `json:"url"`
-	ArchiveFormat string `json:"archive_format"`
-	SHA256        string `json:"sha256"`
-}
-
-func FetchManifest(
-	ctx context.Context,
-	client *http.Client,
-	url string,
-) (*Manifest, error) {
-	req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
-	if err != nil {
-		return nil, err
-	}
-
-	resp, err := client.Do(req)
-	if err != nil {
-		return nil, err
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("update manifest returned HTTP %d", resp.StatusCode)
-	}
-
-	var manifest Manifest
-	if err := json.NewDecoder(io.LimitReader(resp.Body, maxManifestSize)).Decode(&manifest); err != nil {
-		return nil, err
-	}
-	if err := manifest.Validate(); err != nil {
-		return nil, err
-	}
-
-	return &manifest, nil
-}
-
-func (m *Manifest) Validate() error {
-	if m == nil {
-		return fmt.Errorf("update manifest is nil")
-	}
-	if !IsReleaseVersion(m.Version) {
-		return fmt.Errorf("invalid manifest version %q", m.Version)
-	}
-	if len(m.Assets) == 0 {
-		return fmt.Errorf("update manifest has no assets")
-	}
-
-	for key, asset := range m.Assets {
-		if asset.URL == "" {
-			return fmt.Errorf("asset %q is missing url", key)
-		}
-		if asset.SHA256 == "" {
-			return fmt.Errorf("asset %q is missing sha256", key)
-		}
-		if asset.ArchiveFormat != "tar.gz" && asset.ArchiveFormat != "zip" {
-			return fmt.Errorf("asset %q has unsupported archive format %q", key, asset.ArchiveFormat)
-		}
-	}
-
-	return nil
-}
-
-func NormalizeVersion(version string) string {
-	value := strings.TrimSpace(version)
-	if value == "" {
-		return ""
-	}
-	if !strings.HasPrefix(value, "v") {
-		value = "v" + value
-	}
-	return semver.Canonical(value)
-}
-
-func IsReleaseVersion(version string) bool {
-	return NormalizeVersion(version) != ""
-}
-
-func CompareVersions(current, latest string) (int, error) {
-	normalizedCurrent := NormalizeVersion(current)
-	if normalizedCurrent == "" {
-		return 0, fmt.Errorf("invalid current version %q", current)
-	}
-
-	normalizedLatest := NormalizeVersion(latest)
-	if normalizedLatest == "" {
-		return 0, fmt.Errorf("invalid latest version %q", latest)
-	}
-
-	return semver.Compare(normalizedCurrent, normalizedLatest), nil
-}
-
-func PlatformKey(goos, goarch string) (string, error) {
-	switch goos {
-	case "darwin", "linux", "windows":
-	default:
-		return "", fmt.Errorf("unsupported os %q", goos)
-	}
-
-	switch goarch {
-	case "amd64", "arm64":
-	default:
-		return "", fmt.Errorf("unsupported arch %q", goarch)
-	}
-
-	return goos + "/" + goarch, nil
-}
-
-func SelectAsset(manifest *Manifest, goos, goarch string) (Asset, error) {
-	key, err := PlatformKey(goos, goarch)
-	if err != nil {
-		return Asset{}, err
-	}
-
-	asset, ok := manifest.Assets[key]
-	if !ok {
-		return Asset{}, fmt.Errorf("no update asset for %s", key)
-	}
-
-	return asset, nil
-}

packages/browseros-agent/apps/cli/update/manifest_test.go

@@ -1,102 +0,0 @@
-package update
-
-import (
-	"context"
-	"net/http"
-	"net/http/httptest"
-	"strings"
-	"testing"
-)
-
-func TestNormalizeVersion(t *testing.T) {
-	if got := NormalizeVersion("1.2.3"); got != "v1.2.3" {
-		t.Fatalf("NormalizeVersion() = %q, want %q", got, "v1.2.3")
-	}
-	if got := NormalizeVersion("dev"); got != "" {
-		t.Fatalf("NormalizeVersion(dev) = %q, want empty", got)
-	}
-}
-
-func TestCompareVersions(t *testing.T) {
-	got, err := CompareVersions("1.2.3", "1.3.0")
-	if err != nil {
-		t.Fatalf("CompareVersions() error = %v", err)
-	}
-	if got >= 0 {
-		t.Fatalf("CompareVersions() = %d, want < 0", got)
-	}
-}
-
-func TestSelectAsset(t *testing.T) {
-	manifest := &Manifest{
-		Version: "1.2.3",
-		Assets: map[string]Asset{
-			"darwin/arm64": {
-				URL:           "https://cdn.example.com/cli/v1.2.3/browseros-cli.tar.gz",
-				ArchiveFormat: "tar.gz",
-				SHA256:        "abc",
-			},
-		},
-	}
-
-	asset, err := SelectAsset(manifest, "darwin", "arm64")
-	if err != nil {
-		t.Fatalf("SelectAsset() error = %v", err)
-	}
-	if asset.URL == "" {
-		t.Fatal("SelectAsset() returned empty URL")
-	}
-}
-
-func TestFetchManifest(t *testing.T) {
-	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "application/json")
-		_, _ = w.Write([]byte(`{
-			"version":"1.2.3",
-			"published_at":"2026-03-27T19:00:00Z",
-			"tag":"browseros-cli-v1.2.3",
-			"assets":{
-				"darwin/arm64":{
-					"filename":"browseros-cli_1.2.3_darwin_arm64.tar.gz",
-					"url":"https://cdn.example.com/cli/v1.2.3/browseros-cli_1.2.3_darwin_arm64.tar.gz",
-					"archive_format":"tar.gz",
-					"sha256":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
-				}
-			}
-		}`))
-	}))
-	defer server.Close()
-
-	manifest, err := FetchManifest(context.Background(), server.Client(), server.URL)
-	if err != nil {
-		t.Fatalf("FetchManifest() error = %v", err)
-	}
-	if manifest.Version != "1.2.3" {
-		t.Fatalf("FetchManifest() version = %q, want %q", manifest.Version, "1.2.3")
-	}
-}
-
-func TestFetchManifestRejectsOversizedResponse(t *testing.T) {
-	hugeName := strings.Repeat("a", maxManifestSize)
-	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "application/json")
-		_, _ = w.Write([]byte(`{
-			"version":"1.2.3",
-			"published_at":"2026-03-27T19:00:00Z",
-			"tag":"browseros-cli-v1.2.3",
-			"assets":{
-				"darwin/arm64":{
-					"filename":"` + hugeName + `",
-					"url":"https://cdn.example.com/cli/v1.2.3/browseros-cli_1.2.3_darwin_arm64.tar.gz",
-					"archive_format":"tar.gz",
-					"sha256":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
-				}
-			}
-		}`))
-	}))
-	defer server.Close()
-
-	if _, err := FetchManifest(context.Background(), server.Client(), server.URL); err == nil {
-		t.Fatal("FetchManifest() error = nil, want oversized response error")
-	}
-}

packages/browseros-agent/apps/cli/update/state.go

@@ -1,80 +0,0 @@
-package update
-
-import (
-	"encoding/json"
-	"os"
-	"path/filepath"
-	"time"
-
-	"browseros-cli/config"
-)
-
-type State struct {
-	LastCheckedAt     time.Time `json:"last_checked_at"`
-	LatestVersion     string    `json:"latest_version,omitempty"`
-	LatestPublishedAt string    `json:"latest_published_at,omitempty"`
-	AssetURL          string    `json:"asset_url,omitempty"`
-	CheckError        string    `json:"check_error,omitempty"`
-}
-
-func StatePath() string {
-	return filepath.Join(config.Dir(), "update-state.json")
-}
-
-func LoadState() (*State, error) {
-	data, err := os.ReadFile(StatePath())
-	if err != nil {
-		if os.IsNotExist(err) {
-			return &State{}, nil
-		}
-		return nil, err
-	}
-
-	var state State
-	if err := json.Unmarshal(data, &state); err != nil {
-		return nil, err
-	}
-
-	return &state, nil
-}
-
-func SaveState(state *State) error {
-	if state == nil {
-		state = &State{}
-	}
-
-	dir := config.Dir()
-	if err := os.MkdirAll(dir, 0755); err != nil {
-		return err
-	}
-
-	tmpFile, err := os.CreateTemp(dir, "update-state-*.json")
-	if err != nil {
-		return err
-	}
-
-	encoder := json.NewEncoder(tmpFile)
-	encoder.SetIndent("", "  ")
-	if err := encoder.Encode(state); err != nil {
-		tmpFile.Close()
-		os.Remove(tmpFile.Name())
-		return err
-	}
-	if err := tmpFile.Close(); err != nil {
-		os.Remove(tmpFile.Name())
-		return err
-	}
-	if err := os.Rename(tmpFile.Name(), StatePath()); err != nil {
-		os.Remove(tmpFile.Name())
-		return err
-	}
-
-	return nil
-}
-
-func (s *State) IsStale(now time.Time, ttl time.Duration) bool {
-	if s == nil || s.LastCheckedAt.IsZero() {
-		return true
-	}
-	return now.Sub(s.LastCheckedAt) >= ttl
-}

packages/browseros-agent/apps/cli/update/state_test.go

@@ -1,54 +0,0 @@
-package update
-
-import (
-	"path/filepath"
-	"testing"
-	"time"
-)
-
-func TestLoadStateMissing(t *testing.T) {
-	configRoot := t.TempDir()
-	t.Setenv("XDG_CONFIG_HOME", configRoot)
-
-	state, err := LoadState()
-	if err != nil {
-		t.Fatalf("LoadState() error = %v", err)
-	}
-	if state == nil {
-		t.Fatal("LoadState() returned nil state")
-	}
-}
-
-func TestSaveStateRoundTrip(t *testing.T) {
-	configRoot := t.TempDir()
-	t.Setenv("XDG_CONFIG_HOME", configRoot)
-
-	want := &State{
-		LastCheckedAt:     time.Unix(100, 0).UTC(),
-		LatestVersion:     "1.2.3",
-		LatestPublishedAt: "2026-03-27T19:00:00Z",
-		AssetURL:          "https://cdn.example.com/cli/v1.2.3/browseros-cli.tar.gz",
-	}
-	if err := SaveState(want); err != nil {
-		t.Fatalf("SaveState() error = %v", err)
-	}
-
-	got, err := LoadState()
-	if err != nil {
-		t.Fatalf("LoadState() error = %v", err)
-	}
-	if got.LatestVersion != want.LatestVersion {
-		t.Fatalf("LatestVersion = %q, want %q", got.LatestVersion, want.LatestVersion)
-	}
-	if StatePath() != filepath.Join(configRoot, "browseros-cli", "update-state.json") {
-		t.Fatalf("StatePath() = %q", StatePath())
-	}
-}
-
-func TestStateIsStale(t *testing.T) {
-	now := time.Unix(200, 0).UTC()
-	state := &State{LastCheckedAt: time.Unix(0, 0).UTC()}
-	if !state.IsStale(now, time.Minute) {
-		t.Fatal("IsStale() = false, want true")
-	}
-}

packages/browseros-agent/apps/controller-ext/.gitignore

@@ -0,0 +1,32 @@
+# Dependencies
+node_modules/
+
+# Build output
+dist/
+
+# Build unpublished docs
+docs/
+
+# TypeScript
+*.tsbuildinfo
+
+# IDE
+.vscode/
+.idea/
+*.swp
+*.swo
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Logs
+*.log
+npm-debug.log*
+
+# Environment
+.env
+.env.local
+
+# Claude
+.claude

packages/browseros-agent/apps/controller-ext/README.md

@@ -0,0 +1,430 @@
+# BrowserOS Controller
+
+WebSocket-based Chrome Extension that exposes browser automation APIs for remote control.
+
+**⚠️ IMPORTANT:** This extension ONLY works in **BrowserOS Chrome**, not regular Chrome!
+
+---
+
+## 🚀 Quick Start
+
+### 1. Build the Extension
+
+```bash
+npm install
+npm run build
+```
+
+### 2. Load Extension in BrowserOS Chrome
+
+1. Open BrowserOS Chrome
+2. Go to `chrome://extensions/`
+3. Enable **"Developer mode"** (top-right toggle)
+4. Click **"Load unpacked"**
+5. Select the `dist/` folder
+6. Verify extension is loaded (you should see "BrowserOS Controller")
+
+### 3. Test the Extension
+
+```bash
+npm test
+```
+
+This starts an interactive test client. You should see:
+
+```
+🚀 Starting BrowserOS Controller Test Client
+──────────────────────────────────────────────────────────
+
+WebSocket Server Started
+Listening on: ws://localhost:9224/controller
+Waiting for extension to connect...
+
+✅ Extension connected!
+
+Running Diagnostic Test
+============================================================
+
+📤 Sending: checkBrowserOS
+   Request ID: test-1729012345678
+
+📨 Response: test-1729012345678
+   Status: ✅ SUCCESS
+   Data: {
+     "available": true,
+     "apis": [
+       "captureScreenshot",
+       "clear",
+       "click",
+       ...
+     ]
+   }
+```
+
+**If you see "available": true**, you're all set! 🎉
+
+**If you see "available": false**, you're not using BrowserOS Chrome.
+
+---
+
+## ⚙️ Configuration
+
+The extension can be configured using environment variables. This is optional - sensible defaults are provided.
+
+### Environment Variables
+
+Create a `.env` file in the project root to customize configuration:
+
+```bash
+# Copy the example file
+cp .env.example .env
+
+# Edit .env with your values
+```
+
+### Available Configuration Options
+
+#### WebSocket Configuration
+
+```bash
+WEBSOCKET_PROTOCOL=ws          # ws or wss (default: ws)
+WEBSOCKET_HOST=localhost        # Server host (default: localhost)
+WEBSOCKET_PORT=9224            # Server port (default: 9224)
+WEBSOCKET_PATH=/controller     # Server path (default: /controller)
+```
+
+#### Connection Settings
+
+```bash
+WEBSOCKET_RECONNECT_DELAY=1000              # Initial reconnect delay in ms (default: 1000)
+WEBSOCKET_MAX_RECONNECT_DELAY=30000         # Max reconnect delay in ms (default: 30000)
+WEBSOCKET_RECONNECT_MULTIPLIER=1.5          # Exponential backoff multiplier (default: 1.5)
+WEBSOCKET_MAX_RECONNECT_ATTEMPTS=0          # Max reconnect attempts, 0 = infinite (default: 0)
+WEBSOCKET_HEARTBEAT_INTERVAL=30000          # Heartbeat interval in ms (default: 30000)
+WEBSOCKET_HEARTBEAT_TIMEOUT=5000            # Heartbeat timeout in ms (default: 5000)
+WEBSOCKET_CONNECTION_TIMEOUT=10000          # Connection timeout in ms (default: 10000)
+WEBSOCKET_REQUEST_TIMEOUT=30000             # Request timeout in ms (default: 30000)
+```
+
+#### Concurrency Settings
+
+```bash
+CONCURRENCY_MAX_CONCURRENT=100     # Max concurrent requests (default: 100)
+CONCURRENCY_MAX_QUEUE_SIZE=1000    # Max queued requests (default: 1000)
+```
+
+#### Logging Settings
+
+```bash
+LOGGING_ENABLED=true                       # Enable/disable logging (default: true)
+LOGGING_LEVEL=info                         # Log level: debug, info, warn, error (default: info)
+LOGGING_PREFIX=[BrowserOS Controller]      # Log message prefix (default: [BrowserOS Controller])
+```
+
+### Example: Custom Port Configuration
+
+If you want to use a different port (e.g., 8080):
+
+```bash
+# .env
+WEBSOCKET_PORT=8080
+```
+
+Then rebuild the extension:
+
+```bash
+npm run build
+```
+
+The extension will now connect to `ws://localhost:8080/controller` instead of the default port 9224.
+
+---
+
+## 📖 Architecture
+
+See [ARCHITECTURE.md](./ARCHITECTURE.md) for complete system documentation including:
+
+- High-level architecture diagram
+- Request flow (step-by-step)
+- Component details
+- All 14 registered actions
+- WebSocket protocol specification
+- Debugging guide
+
+---
+
+## 🧪 Testing
+
+The test client (`npm test`) provides an interactive menu:
+
+```
+Available Commands:
+
+  Tab Actions:
+  1. getActiveTab       - Get currently active tab
+  2. getTabs            - Get all tabs
+
+  Browser Actions:
+  3. getInteractiveSnapshot  - Get page elements (requires tabId)
+  4. click              - Click element (requires tabId, nodeId)
+  5. inputText          - Type text (requires tabId, nodeId, text)
+  6. captureScreenshot  - Take screenshot (requires tabId)
+
+  Diagnostic:
+  d. checkBrowserOS     - Check if chrome.browserOS is available
+
+  Other:
+  h. Show this menu
+  q. Quit
+```
+
+### Example Usage:
+
+1. Type `1` → Get active tab
+2. Type `d` → Run diagnostic
+3. Type `q` → Quit
+
+---
+
+## 🔧 Development
+
+### Build Commands
+
+```bash
+npm run build      # Production build
+npm run build:dev  # Development build (with source maps)
+npm run watch      # Watch mode for development
+```
+
+### Debug Extension
+
+1. Go to `chrome://extensions/`
+2. Click **"Inspect views service worker"** under "BrowserOS Controller"
+3. Service worker console shows all logs
+
+**Check extension status:**
+
+```javascript
+__browserosController.getStats();
+```
+
+**Expected output:**
+
+```javascript
+{
+  connection: "connected",
+  requests: { inFlight: 0, avgDuration: 0, errorRate: 0, totalRequests: 0 },
+  concurrency: { inFlight: 0, queued: 0, utilization: 0 },
+  validator: { activeIds: 0 },
+  responseQueue: { size: 0 }
+}
+```
+
+**Check registered actions:**
+Look for this log on extension load:
+
+```
+Registered 14 action(s): checkBrowserOS, getActiveTab, getTabs, ...
+```
+
+---
+
+## 📋 Available Actions
+
+| Action                   | Input                             | Output                          | Description                            |
+| ------------------------ | --------------------------------- | ------------------------------- | -------------------------------------- |
+| `checkBrowserOS`         | `{}`                              | `{available, apis}`             | Check if chrome.browserOS is available |
+| `getActiveTab`           | `{}`                              | `{tabId, url, title, windowId}` | Get currently active tab               |
+| `getTabs`                | `{}`                              | `{tabs[]}`                      | Get all open tabs                      |
+| `getInteractiveSnapshot` | `{tabId, options?}`               | `InteractiveSnapshot`           | Get all interactive elements on page   |
+| `click`                  | `{tabId, nodeId}`                 | `{success}`                     | Click element by nodeId                |
+| `inputText`              | `{tabId, nodeId, text}`           | `{success}`                     | Type text into element                 |
+| `clear`                  | `{tabId, nodeId}`                 | `{success}`                     | Clear text from element                |
+| `scrollToNode`           | `{tabId, nodeId}`                 | `{scrolled}`                    | Scroll element into view               |
+| `captureScreenshot`      | `{tabId, size?, showHighlights?}` | `{dataUrl}`                     | Take screenshot                        |
+| `sendKeys`               | `{tabId, keys}`                   | `{success}`                     | Send keyboard keys                     |
+| `getPageLoadStatus`      | `{tabId}`                         | `PageLoadStatus`                | Get page load status                   |
+| `getSnapshot`            | `{tabId, type, options?}`         | `Snapshot`                      | Get text/links snapshot                |
+| `clickCoordinates`       | `{tabId, x, y}`                   | `{success}`                     | Click at coordinates                   |
+| `typeAtCoordinates`      | `{tabId, x, y, text}`             | `{success}`                     | Type at coordinates                    |
+
+---
+
+## 🔌 WebSocket Protocol
+
+**Endpoint:** `ws://localhost:9224/controller`
+
+**Request Format:**
+
+```json
+{
+  "id": "unique-request-id",
+  "action": "click",
+  "payload": {
+    "tabId": 12345,
+    "nodeId": 42
+  }
+}
+```
+
+**Response Format:**
+
+```json
+{
+  "id": "unique-request-id",
+  "ok": true,
+  "data": {
+    "success": true
+  }
+}
+```
+
+**Error Response:**
+
+```json
+{
+  "id": "unique-request-id",
+  "ok": false,
+  "error": "Element not found: nodeId 42"
+}
+```
+
+---
+
+## ⚠️ Common Issues
+
+### Issue 1: "chrome.browserOS is undefined"
+
+**Symptoms:**
+
+- Diagnostic shows `"available": false`
+- All browser actions fail
+
+**Cause:** Not using BrowserOS Chrome
+
+**Solution:**
+
+- Download and use BrowserOS Chrome (not regular Chrome)
+- Verify at `chrome://version` - should show "BrowserOS" in the name
+
+---
+
+### Issue 2: "Port 9224 is already in use"
+
+**Symptoms:**
+
+```
+❌ Fatal Error: Port 9224 is already in use!
+```
+
+**Solution:**
+
+```bash
+lsof -ti:9224 | xargs kill -9
+npm test
+```
+
+---
+
+### Issue 3: Extension Not Connecting
+
+**Symptoms:**
+
+- Test client shows "Waiting for extension to connect..." forever
+- Service worker console shows "Connection timeout"
+
+**Checklist:**
+
+1. ✅ Test server running (`npm test`)
+2. ✅ Extension loaded in BrowserOS Chrome
+3. ✅ Extension enabled (chrome://extensions/)
+4. ✅ Service worker active (not suspended)
+
+**Solution:**
+
+1. Reload extension: chrome://extensions/ → "Reload" button
+2. Restart test server: Ctrl+C, then `npm test`
+
+---
+
+### Issue 4: "Unknown action"
+
+**Symptoms:**
+
+```
+Error: Unknown action: "click". Available actions: getActiveTab, getTabs, ...
+```
+
+**Cause:** Action not registered (extension didn't reload properly)
+
+**Solution:**
+
+1. Toggle extension OFF and ON at chrome://extensions/
+2. Check service worker console for: `Registered 14 action(s): ...`
+
+---
+
+## 📁 Project Structure
+
+```
+browseros-controller/
+├── README.md              # This file
+├── ARCHITECTURE.md        # Complete architecture documentation
+├── .env.example           # Environment variable template
+├── manifest.json          # Extension manifest
+├── package.json           # Node dependencies
+├── webpack.config.js      # Build configuration
+│
+├── src/                   # Source code
+│   ├── background/        # Service worker entry point
+│   ├── actions/           # Action handlers
+│   │   ├── bookmark/      # Bookmark management actions
+│   │   ├── browser/       # Browser interaction actions
+│   │   ├── diagnostics/   # Diagnostic actions
+│   │   ├── history/       # History management actions
+│   │   └── tab/           # Tab management actions
+│   ├── adapters/          # Chrome API wrappers
+│   ├── config/            # Configuration management
+│   │   ├── constants.ts   # Application constants
+│   │   └── environment.ts # Environment variable handling
+│   ├── websocket/         # WebSocket client
+│   ├── utils/             # Utilities
+│   ├── protocol/          # Protocol types
+│   └── types/             # TypeScript definitions
+│
+├── tests/                 # Test files
+│   ├── test-simple.js     # Interactive test client
+│   └── test-auto.js       # Automated test client
+│
+└── dist/                  # Built extension (generated)
+    ├── background.js
+    └── manifest.json
+```
+
+---
+
+## 🔗 Related Projects
+
+- **BrowserOS-agent**: AI agent that uses this controller for browser automation
+- **BrowserOS Chrome**: Custom Chrome build with `chrome.browserOS` APIs
+
+---
+
+## 📄 License
+
+MIT
+
+---
+
+## 🆘 Support
+
+For issues or questions:
+
+1. Check [ARCHITECTURE.md](./ARCHITECTURE.md) for detailed documentation
+2. Review the "Common Issues" section above
+3. Check service worker console for detailed error logs
+4. Verify you're using BrowserOS Chrome (run diagnostic test)
+
+---
+
+**Happy automating! 🚀**

packages/browseros-agent/apps/controller-ext/manifest.json

@@ -0,0 +1,38 @@
+{
+  "manifest_version": 3,
+  "name": "BrowserOS Controller",
+  "version": "1.0.0.8",
+  "description": "BrowserOS API bridge for BrowserOS Server",
+  "key": "MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAhlh9i/c2A3f0PL86hXhGPzguLIOQ+sPf3/Y8RD11gmdvoU6XqnUqv7GgBvm7SW7316uPnS58AYZY13jGtF4rFrscdda5H2CjZrtOyOycmKp2KzibJLwibXNm/JwKhZ3QEfgsW/orh1SMY2kNj62JemkWLcLyn3E1T+KTcTVyFOxiJS3hyQ+Y0/Jp1HOqGh5lYS58YYzwhId5rrJjfL7wFYtALgt2dEA2r7p4qpe+SW0QLA+ayjRAjS+yt+qitR0eWg+XgqcIk1f1KblN8/yDISssSD4LWiPofe5CmJPnqlHIuI0CpgvAFv9dvgR/w8OFkXxK5h06i6saum1xExj+IwIDAQAB",
+  "permissions": [
+    "tabs",
+    "activeTab",
+    "bookmarks",
+    "history",
+    "scripting",
+    "storage",
+    "tabGroups",
+    "webNavigation",
+    "downloads",
+    "browserOS",
+    "alarms"
+  ],
+  "update_url": "https://cdn.browseros.com/extensions/update-manifest.xml",
+  "host_permissions": ["<all_urls>"],
+  "background": {
+    "service_worker": "background.js",
+    "type": "module"
+  },
+  "action": {
+    "default_icon": {
+      "16": "assets/icon16.png",
+      "48": "assets/icon48.png",
+      "128": "assets/icon128.png"
+    }
+  },
+  "icons": {
+    "16": "assets/icon16.png",
+    "48": "assets/icon48.png",
+    "128": "assets/icon128.png"
+  }
+}

packages/browseros-agent/apps/controller-ext/package.json

@@ -0,0 +1,39 @@
+{
+  "name": "browseros-controller",
+  "version": "1.0.0",
+  "description": "Chrome Extension API bridge for BrowserOS Server",
+  "directories": {
+    "doc": "docs"
+  },
+  "scripts": {
+    "build": "webpack --mode production",
+    "build:dev": "webpack --mode development",
+    "watch": "webpack --mode development --watch",
+    "test": "node tests/test-simple.js",
+    "test:auto": "node tests/test-auto.js",
+    "typecheck": "tsc --noEmit"
+  },
+  "keywords": [
+    "browser-automation",
+    "chrome-extension",
+    "browseros"
+  ],
+  "author": "BrowserOS Team",
+  "license": "MIT",
+  "type": "commonjs",
+  "dependencies": {
+    "@browseros/shared": "workspace:*",
+    "zod": "^4.1.12"
+  },
+  "devDependencies": {
+    "@types/chrome": "^0.1.24",
+    "@types/node": "^24.7.1",
+    "copy-webpack-plugin": "^12.0.2",
+    "terser-webpack-plugin": "^5.3.11",
+    "ts-loader": "^9.5.4",
+    "typescript": "^5.9.3",
+    "webpack": "^5.102.1",
+    "webpack-cli": "^6.0.1",
+    "ws": "^8.18.3"
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/ActionHandler.ts

@@ -0,0 +1,106 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+
+import type { ActionResponse } from '@/protocol/types'
+import { ActionResponseSchema } from '@/protocol/types'
+import { logger } from '@/utils/logger'
+
+// Re-export for convenience
+export type { ActionResponse }
+export { ActionResponseSchema }
+
+/**
+ * ActionHandler - Abstract base class for all actions
+ *
+ * Responsibilities:
+ * - Define contract for all actions (must implement inputSchema + execute)
+ * - Validate input using Zod schemas
+ * - Handle validation and execution errors
+ * - Return standardized ActionResponse
+ *
+ * Usage:
+ * class MyAction extends ActionHandler<InputType, OutputType> {
+ *   inputSchema = z.object({ ... });
+ *   async execute(input: InputType): Promise<OutputType> { ... }
+ * }
+ */
+export abstract class ActionHandler<TInput = unknown, TOutput = unknown> {
+  /**
+   * Zod schema for input validation
+   * Must be implemented by concrete actions
+   */
+  abstract readonly inputSchema: z.ZodSchema<TInput>
+
+  /**
+   * Execute the action logic
+   * Must be implemented by concrete actions
+   *
+   * @param input - Validated input (guaranteed to match inputSchema)
+   * @returns Action result
+   */
+  abstract execute(input: TInput): Promise<TOutput>
+
+  /**
+   * Handle request with validation and error handling
+   * Called by ActionRegistry
+   *
+   * Flow:
+   * 1. Validate input with Zod schema
+   * 2. Execute action logic
+   * 3. Return standardized response (ok/error)
+   *
+   * @param payload - Raw payload from request (unvalidated)
+   * @returns Standardized action response
+   */
+  async handle(payload: unknown): Promise<ActionResponse> {
+    const actionName = this.constructor.name
+
+    try {
+      // Step 1: Validate input
+      logger.debug(`[${actionName}] Validating input`)
+      const validatedInput = this.inputSchema.parse(payload)
+
+      // Step 2: Execute action
+      logger.debug(`[${actionName}] Executing action`)
+      const result = await this.execute(validatedInput)
+
+      // Step 3: Return success response
+      logger.debug(`[${actionName}] Action completed successfully`)
+      return { ok: true, data: result }
+    } catch (error) {
+      // Handle validation or execution errors
+      const errorMessage = this._formatError(error)
+      logger.error(`[${actionName}] Action failed: ${errorMessage}`)
+      return { ok: false, error: errorMessage }
+    }
+  }
+
+  /**
+   * Format error for user-friendly response
+   *
+   * @param error - Error from validation or execution
+   * @returns Formatted error message
+   */
+  protected _formatError(error: unknown): string {
+    // Zod validation error
+    if (error instanceof z.ZodError) {
+      const errors = error.issues.map((e: z.ZodIssue) => {
+        const path = e.path.length > 0 ? `${e.path.join('.')}: ` : ''
+        return `${path}${e.message}`
+      })
+      return `Validation error: ${errors.join(', ')}`
+    }
+
+    // Standard Error
+    if (error instanceof Error) {
+      return error.message
+    }
+
+    // Unknown error
+    return String(error)
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/ActionRegistry.ts

@@ -0,0 +1,148 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+
+import { logger } from '@/utils/logger'
+import type { ActionHandler, ActionResponse } from './ActionHandler'
+
+/**
+ * ActionRegistry - Central dispatcher for all actions
+ *
+ * Responsibilities:
+ * - Register action handlers by name
+ * - Dispatch requests to correct handler
+ * - Return error for unknown actions
+ * - Provide introspection (list available actions)
+ *
+ * Usage:
+ * const registry = new ActionRegistry();
+ * registry.register('getActiveTab', new GetActiveTabAction());
+ * const response = await registry.dispatch('getActiveTab', {});
+ */
+export class ActionRegistry {
+  private handlers = new Map<string, ActionHandler>()
+
+  /**
+   * Register an action handler
+   *
+   * @param actionName - Unique action name (e.g., "getActiveTab")
+   * @param handler - Action handler instance
+   */
+  register(actionName: string, handler: ActionHandler): void {
+    if (this.handlers.has(actionName)) {
+      logger.warn(
+        `[ActionRegistry] Action "${actionName}" already registered, overwriting`,
+      )
+    }
+
+    this.handlers.set(actionName, handler)
+    logger.info(`[ActionRegistry] Registered action: ${actionName}`)
+  }
+
+  /**
+   * Dispatch request to appropriate action handler
+   *
+   * Flow:
+   * 1. Find handler for action name
+   * 2. If not found, return error
+   * 3. If found, delegate to handler.handle()
+   * 4. Handler validates input and executes
+   * 5. Return result
+   *
+   * @param actionName - Action to execute
+   * @param payload - Action payload (unvalidated)
+   * @returns Action response
+   */
+  async dispatch(
+    actionName: string,
+    payload: unknown,
+  ): Promise<ActionResponse> {
+    logger.debug(`[ActionRegistry] Dispatching action: ${actionName}`)
+
+    // Check if action exists
+    const handler = this.handlers.get(actionName)
+
+    if (!handler) {
+      const availableActions = Array.from(this.handlers.keys()).join(', ')
+      const errorMessage = `Unknown action: "${actionName}". Available actions: ${availableActions || 'none'}`
+      logger.error(`[ActionRegistry] ${errorMessage}`)
+      return {
+        ok: false,
+        error: errorMessage,
+      }
+    }
+
+    // Delegate to handler
+    try {
+      const response = await handler.handle(payload)
+      logger.debug(
+        `[ActionRegistry] Action "${actionName}" ${response.ok ? 'succeeded' : 'failed'}`,
+      )
+      return response
+    } catch (error) {
+      // Catch any unexpected errors from handler
+      const errorMessage =
+        error instanceof Error ? error.message : String(error)
+      logger.error(
+        `[ActionRegistry] Unexpected error in "${actionName}": ${errorMessage}`,
+      )
+      return {
+        ok: false,
+        error: `Action execution failed: ${errorMessage}`,
+      }
+    }
+  }
+
+  /**
+   * Get list of registered action names
+   *
+   * @returns Array of action names
+   */
+  getAvailableActions(): string[] {
+    return Array.from(this.handlers.keys())
+  }
+
+  /**
+   * Check if action is registered
+   *
+   * @param actionName - Action name to check
+   * @returns True if action exists
+   */
+  hasAction(actionName: string): boolean {
+    return this.handlers.has(actionName)
+  }
+
+  /**
+   * Get number of registered actions
+   *
+   * @returns Count of registered actions
+   */
+  getActionCount(): number {
+    return this.handlers.size
+  }
+
+  /**
+   * Unregister an action (useful for testing)
+   *
+   * @param actionName - Action to remove
+   * @returns True if action was removed
+   */
+  unregister(actionName: string): boolean {
+    const removed = this.handlers.delete(actionName)
+    if (removed) {
+      logger.info(`[ActionRegistry] Unregistered action: ${actionName}`)
+    }
+    return removed
+  }
+
+  /**
+   * Clear all registered actions (useful for testing)
+   */
+  clear(): void {
+    const count = this.handlers.size
+    this.handlers.clear()
+    logger.info(`[ActionRegistry] Cleared ${count} registered actions`)
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/CreateBookmarkAction.ts

@@ -0,0 +1,81 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const CreateBookmarkInputSchema = z.object({
+  title: z.string().describe('Bookmark title'),
+  url: z.string().url().describe('Bookmark URL'),
+  parentId: z
+    .string()
+    .optional()
+    .describe('Parent folder ID (optional, defaults to "Other Bookmarks")'),
+})
+
+// Output schema
+const CreateBookmarkOutputSchema = z.object({
+  id: z.string().describe('Created bookmark ID'),
+  title: z.string().describe('Bookmark title'),
+  url: z.string().describe('Bookmark URL'),
+  dateAdded: z
+    .number()
+    .optional()
+    .describe('Timestamp when bookmark was created'),
+})
+
+type CreateBookmarkInput = z.infer<typeof CreateBookmarkInputSchema>
+type CreateBookmarkOutput = z.infer<typeof CreateBookmarkOutputSchema>
+
+/**
+ * CreateBookmarkAction - Create a new bookmark
+ *
+ * Creates a bookmark with the specified title and URL.
+ *
+ * Input:
+ * - title: Display title for the bookmark
+ * - url: Full URL to bookmark
+ * - parentId (optional): Parent folder ID
+ *
+ * Output:
+ * - id: Created bookmark ID
+ * - title: Bookmark title
+ * - url: Bookmark URL
+ * - dateAdded: Creation timestamp
+ *
+ * Usage:
+ * Create a bookmark in the default location (Other Bookmarks).
+ *
+ * Example:
+ * {
+ *   "title": "Google",
+ *   "url": "https://www.google.com"
+ * }
+ * // Returns: { id: "123", title: "Google", url: "https://www.google.com", dateAdded: 1729012345678 }
+ */
+export class CreateBookmarkAction extends ActionHandler<
+  CreateBookmarkInput,
+  CreateBookmarkOutput
+> {
+  readonly inputSchema = CreateBookmarkInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(input: CreateBookmarkInput): Promise<CreateBookmarkOutput> {
+    const created = await this.bookmarkAdapter.createBookmark({
+      title: input.title,
+      url: input.url,
+      parentId: input.parentId,
+    })
+
+    return {
+      id: created.id,
+      title: created.title,
+      url: created.url || '',
+      dateAdded: created.dateAdded,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/CreateBookmarkFolderAction.ts

@@ -0,0 +1,52 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+const CreateBookmarkFolderInputSchema = z.object({
+  title: z.string().describe('Folder name'),
+  parentId: z
+    .string()
+    .optional()
+    .describe('Parent folder ID (defaults to "1" = Bookmarks Bar)'),
+})
+
+const CreateBookmarkFolderOutputSchema = z.object({
+  id: z.string().describe('Created folder ID'),
+  title: z.string().describe('Folder name'),
+  parentId: z.string().optional().describe('Parent folder ID'),
+  dateAdded: z.number().optional().describe('Creation timestamp'),
+})
+
+type CreateBookmarkFolderInput = z.infer<typeof CreateBookmarkFolderInputSchema>
+type CreateBookmarkFolderOutput = z.infer<
+  typeof CreateBookmarkFolderOutputSchema
+>
+
+export class CreateBookmarkFolderAction extends ActionHandler<
+  CreateBookmarkFolderInput,
+  CreateBookmarkFolderOutput
+> {
+  readonly inputSchema = CreateBookmarkFolderInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(
+    input: CreateBookmarkFolderInput,
+  ): Promise<CreateBookmarkFolderOutput> {
+    const created = await this.bookmarkAdapter.createBookmarkFolder({
+      title: input.title,
+      parentId: input.parentId,
+    })
+
+    return {
+      id: created.id,
+      title: created.title,
+      parentId: created.parentId,
+      dateAdded: created.dateAdded,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/GetBookmarkChildrenAction.ts

@@ -0,0 +1,59 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+const GetBookmarkChildrenInputSchema = z.object({
+  folderId: z.string().describe('Folder ID to get children from'),
+})
+
+const GetBookmarkChildrenOutputSchema = z.object({
+  children: z.array(
+    z.object({
+      id: z.string(),
+      title: z.string(),
+      url: z.string().optional(),
+      parentId: z.string().optional(),
+      dateAdded: z.number().optional(),
+      isFolder: z.boolean(),
+    }),
+  ),
+  count: z.number(),
+})
+
+type GetBookmarkChildrenInput = z.infer<typeof GetBookmarkChildrenInputSchema>
+type GetBookmarkChildrenOutput = z.infer<typeof GetBookmarkChildrenOutputSchema>
+
+export class GetBookmarkChildrenAction extends ActionHandler<
+  GetBookmarkChildrenInput,
+  GetBookmarkChildrenOutput
+> {
+  readonly inputSchema = GetBookmarkChildrenInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(
+    input: GetBookmarkChildrenInput,
+  ): Promise<GetBookmarkChildrenOutput> {
+    const results = await this.bookmarkAdapter.getBookmarkChildren(
+      input.folderId,
+    )
+
+    const children = results.map((node) => ({
+      id: node.id,
+      title: node.title,
+      url: node.url,
+      parentId: node.parentId,
+      dateAdded: node.dateAdded,
+      isFolder: !node.url,
+    }))
+
+    return {
+      children,
+      count: children.length,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/GetBookmarksAction.ts

@@ -0,0 +1,111 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const GetBookmarksInputSchema = z.object({
+  query: z
+    .string()
+    .optional()
+    .describe(
+      'Search query to filter bookmarks (optional, returns all if not provided)',
+    ),
+  limit: z
+    .number()
+    .int()
+    .positive()
+    .optional()
+    .default(20)
+    .describe('Maximum number of results (default: 20)'),
+  recent: z
+    .boolean()
+    .optional()
+    .default(false)
+    .describe('Get recent bookmarks instead of searching'),
+})
+
+// Output schema
+const GetBookmarksOutputSchema = z.object({
+  bookmarks: z.array(
+    z.object({
+      id: z.string(),
+      title: z.string(),
+      url: z.string().optional(),
+      dateAdded: z.number().optional(),
+      parentId: z.string().optional(),
+    }),
+  ),
+  count: z.number(),
+})
+
+type GetBookmarksInput = z.infer<typeof GetBookmarksInputSchema>
+type GetBookmarksOutput = z.infer<typeof GetBookmarksOutputSchema>
+
+/**
+ * GetBookmarksAction - Get or search bookmarks
+ *
+ * Retrieves bookmarks with optional filtering.
+ *
+ * Input:
+ * - query (optional): Search query to match title or URL
+ * - limit (optional): Maximum results (default: 20)
+ * - recent (optional): Get recent bookmarks instead (default: false)
+ *
+ * Output:
+ * - bookmarks: Array of bookmark objects
+ * - count: Number of bookmarks returned
+ *
+ * Usage:
+ * - Get recent: { "recent": true }
+ * - Search: { "query": "github" }
+ * - Get all (limited): { "limit": 50 }
+ *
+ * Example:
+ * {
+ *   "query": "google",
+ *   "limit": 10
+ * }
+ * // Returns: { bookmarks: [{id: "1", title: "Google", url: "https://google.com"}], count: 1 }
+ */
+export class GetBookmarksAction extends ActionHandler<
+  GetBookmarksInput,
+  GetBookmarksOutput
+> {
+  readonly inputSchema = GetBookmarksInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(input: GetBookmarksInput): Promise<GetBookmarksOutput> {
+    let results: chrome.bookmarks.BookmarkTreeNode[]
+
+    if (input.recent) {
+      // Get recent bookmarks
+      results = await this.bookmarkAdapter.getRecentBookmarks(input.limit)
+    } else if (input.query) {
+      // Search bookmarks
+      results = await this.bookmarkAdapter.searchBookmarks(input.query)
+      results = results.slice(0, input.limit)
+    } else {
+      // Get recent by default
+      results = await this.bookmarkAdapter.getRecentBookmarks(input.limit)
+    }
+
+    // Map to output format
+    const bookmarks = results.map((b) => ({
+      id: b.id,
+      title: b.title,
+      url: b.url,
+      dateAdded: b.dateAdded,
+      parentId: b.parentId,
+    }))
+
+    return {
+      bookmarks,
+      count: bookmarks.length,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/MoveBookmarkAction.ts

@@ -0,0 +1,49 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+const MoveBookmarkInputSchema = z.object({
+  id: z.string().describe('Bookmark or folder ID to move'),
+  parentId: z.string().optional().describe('New parent folder ID'),
+  index: z.number().int().min(0).optional().describe('Position within parent'),
+})
+
+const MoveBookmarkOutputSchema = z.object({
+  id: z.string().describe('Moved bookmark ID'),
+  title: z.string().describe('Bookmark title'),
+  url: z.string().optional().describe('Bookmark URL (undefined if folder)'),
+  parentId: z.string().optional().describe('New parent folder ID'),
+  index: z.number().optional().describe('New position within parent'),
+})
+
+type MoveBookmarkInput = z.infer<typeof MoveBookmarkInputSchema>
+type MoveBookmarkOutput = z.infer<typeof MoveBookmarkOutputSchema>
+
+export class MoveBookmarkAction extends ActionHandler<
+  MoveBookmarkInput,
+  MoveBookmarkOutput
+> {
+  readonly inputSchema = MoveBookmarkInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(input: MoveBookmarkInput): Promise<MoveBookmarkOutput> {
+    const destination: { parentId?: string; index?: number } = {}
+    if (input.parentId !== undefined) destination.parentId = input.parentId
+    if (input.index !== undefined) destination.index = input.index
+
+    const moved = await this.bookmarkAdapter.moveBookmark(input.id, destination)
+
+    return {
+      id: moved.id,
+      title: moved.title,
+      url: moved.url,
+      parentId: moved.parentId,
+      index: moved.index,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/RemoveBookmarkAction.ts

@@ -0,0 +1,62 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const RemoveBookmarkInputSchema = z.object({
+  id: z.string().describe('Bookmark ID to remove'),
+})
+
+// Output schema
+const RemoveBookmarkOutputSchema = z.object({
+  success: z
+    .boolean()
+    .describe('Whether the bookmark was successfully removed'),
+  message: z.string().describe('Confirmation message'),
+})
+
+type RemoveBookmarkInput = z.infer<typeof RemoveBookmarkInputSchema>
+type RemoveBookmarkOutput = z.infer<typeof RemoveBookmarkOutputSchema>
+
+/**
+ * RemoveBookmarkAction - Remove a bookmark
+ *
+ * Deletes a bookmark by its ID.
+ *
+ * Input:
+ * - id: Bookmark ID to remove
+ *
+ * Output:
+ * - success: true if removed
+ * - message: Confirmation message
+ *
+ * Usage:
+ * Get the bookmark ID from getBookmarks first, then remove it.
+ *
+ * Example:
+ * {
+ *   "id": "123"
+ * }
+ * // Returns: { success: true, message: "Removed bookmark 123" }
+ */
+export class RemoveBookmarkAction extends ActionHandler<
+  RemoveBookmarkInput,
+  RemoveBookmarkOutput
+> {
+  readonly inputSchema = RemoveBookmarkInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(input: RemoveBookmarkInput): Promise<RemoveBookmarkOutput> {
+    await this.bookmarkAdapter.removeBookmark(input.id)
+
+    return {
+      success: true,
+      message: `Removed bookmark ${input.id}`,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/RemoveBookmarkTreeAction.ts

@@ -0,0 +1,48 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+const RemoveBookmarkTreeInputSchema = z.object({
+  id: z.string().describe('Folder ID to remove'),
+  confirm: z.boolean().describe('Must be true to confirm recursive deletion'),
+})
+
+const RemoveBookmarkTreeOutputSchema = z.object({
+  success: z.boolean().describe('Whether the folder was removed'),
+  message: z.string().describe('Result message'),
+})
+
+type RemoveBookmarkTreeInput = z.infer<typeof RemoveBookmarkTreeInputSchema>
+type RemoveBookmarkTreeOutput = z.infer<typeof RemoveBookmarkTreeOutputSchema>
+
+export class RemoveBookmarkTreeAction extends ActionHandler<
+  RemoveBookmarkTreeInput,
+  RemoveBookmarkTreeOutput
+> {
+  readonly inputSchema = RemoveBookmarkTreeInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(
+    input: RemoveBookmarkTreeInput,
+  ): Promise<RemoveBookmarkTreeOutput> {
+    if (input.confirm !== true) {
+      return {
+        success: false,
+        message:
+          'Recursive deletion requires confirm: true. This will permanently delete the folder and all its contents.',
+      }
+    }
+
+    await this.bookmarkAdapter.removeBookmarkTree(input.id)
+
+    return {
+      success: true,
+      message: `Removed folder ${input.id} and all its contents`,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/bookmark/UpdateBookmarkAction.ts

@@ -0,0 +1,82 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BookmarkAdapter } from '@/adapters/BookmarkAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const UpdateBookmarkInputSchema = z.object({
+  id: z.string().describe('Bookmark ID to update'),
+  title: z.string().optional().describe('New bookmark title'),
+  url: z.string().url().optional().describe('New bookmark URL'),
+})
+
+// Output schema
+const UpdateBookmarkOutputSchema = z.object({
+  id: z.string().describe('Bookmark ID'),
+  title: z.string().describe('Updated bookmark title'),
+  url: z.string().optional().describe('Updated bookmark URL'),
+})
+
+type UpdateBookmarkInput = z.infer<typeof UpdateBookmarkInputSchema>
+type UpdateBookmarkOutput = z.infer<typeof UpdateBookmarkOutputSchema>
+
+/**
+ * UpdateBookmarkAction - Update a bookmark's title or URL
+ *
+ * Updates an existing bookmark with new title and/or URL.
+ *
+ * Input:
+ * - id: Bookmark ID to update
+ * - title (optional): New title for the bookmark
+ * - url (optional): New URL for the bookmark
+ *
+ * Output:
+ * - id: Bookmark ID
+ * - title: Updated title
+ * - url: Updated URL
+ *
+ * Usage:
+ * Update a bookmark's title or URL (at least one must be provided).
+ *
+ * Example:
+ * {
+ *   "id": "123",
+ *   "title": "New Title",
+ *   "url": "https://www.example.com"
+ * }
+ * // Returns: { id: "123", title: "New Title", url: "https://www.example.com" }
+ */
+export class UpdateBookmarkAction extends ActionHandler<
+  UpdateBookmarkInput,
+  UpdateBookmarkOutput
+> {
+  readonly inputSchema = UpdateBookmarkInputSchema
+  private bookmarkAdapter = new BookmarkAdapter()
+
+  async execute(input: UpdateBookmarkInput): Promise<UpdateBookmarkOutput> {
+    const changes: { title?: string; url?: string } = {}
+
+    if (input.title !== undefined) {
+      changes.title = input.title
+    }
+    if (input.url !== undefined) {
+      changes.url = input.url
+    }
+
+    if (Object.keys(changes).length === 0) {
+      throw new Error('At least one of title or url must be provided')
+    }
+
+    const updated = await this.bookmarkAdapter.updateBookmark(input.id, changes)
+
+    return {
+      id: updated.id,
+      title: updated.title,
+      url: updated.url,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/CaptureScreenshotAction.ts

@@ -0,0 +1,79 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import {
+  BrowserOSAdapter,
+  type ScreenshotSizeKey,
+} from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const CaptureScreenshotInputSchema = z.object({
+  tabId: z.number().describe('The tab ID to capture'),
+  size: z
+    .enum(['small', 'medium', 'large'])
+    .optional()
+    .default('medium')
+    .describe('Screenshot size preset (default: medium)'),
+  showHighlights: z
+    .boolean()
+    .optional()
+    .default(true)
+    .describe('Show element highlights (default: true)'),
+  width: z.number().optional().describe('Exact width in pixels'),
+  height: z.number().optional().describe('Exact height in pixels'),
+})
+
+// Output schema
+const CaptureScreenshotOutputSchema = z.object({
+  dataUrl: z.string().describe('Base64-encoded PNG data URL'),
+})
+
+type CaptureScreenshotInput = z.infer<typeof CaptureScreenshotInputSchema>
+type CaptureScreenshotOutput = z.infer<typeof CaptureScreenshotOutputSchema>
+
+/**
+ * CaptureScreenshotAction - Capture a screenshot of the page
+ *
+ * Captures a screenshot with configurable size and options.
+ *
+ * Size Options:
+ * - small (512px): Low detail, minimal tokens
+ * - medium (768px): Balanced quality/tokens (default)
+ * - large (1028px): High detail, maximum tokens
+ *
+ * Or specify exact dimensions with width/height.
+ *
+ * Returns:
+ * - dataUrl: PNG image as base64 data URL (data:image/png;base64,...)
+ *
+ * Usage:
+ * 1. For AI vision models: use 'medium' or 'large'
+ * 2. For debugging: use 'small'
+ * 3. For exact size: specify width and height
+ *
+ * Used by: ScreenshotTool, VisualClick, VisualType
+ */
+export class CaptureScreenshotAction extends ActionHandler<
+  CaptureScreenshotInput,
+  CaptureScreenshotOutput
+> {
+  readonly inputSchema = CaptureScreenshotInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(
+    input: CaptureScreenshotInput,
+  ): Promise<CaptureScreenshotOutput> {
+    const dataUrl = await this.browserOSAdapter.captureScreenshot(
+      input.tabId,
+      input.size as ScreenshotSizeKey | undefined,
+      input.showHighlights,
+      input.width,
+      input.height,
+    )
+    return { dataUrl }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/CaptureScreenshotPointerAction.ts

@@ -0,0 +1,124 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import {
+  BrowserOSAdapter,
+  type ScreenshotSizeKey,
+} from '@/adapters/BrowserOSAdapter'
+import { logger } from '@/utils/logger'
+import { PointerOverlay } from '@/utils/PointerOverlay'
+import { SnapshotCache } from '@/utils/SnapshotCache'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const CaptureScreenshotPointerInputSchema = z.object({
+  tabId: z.number().describe('The tab ID to capture'),
+  nodeId: z
+    .number()
+    .int()
+    .positive()
+    .describe('The nodeId to show pointer over'),
+  size: z
+    .enum(['small', 'medium', 'large'])
+    .optional()
+    .default('medium')
+    .describe('Screenshot size preset (default: medium)'),
+  pointerLabel: z
+    .string()
+    .optional()
+    .describe('Optional label to show with pointer (e.g., "Click", "Type")'),
+})
+
+// Output schema
+const CaptureScreenshotPointerOutputSchema = z.object({
+  dataUrl: z.string().describe('Base64-encoded PNG data URL'),
+  pointerPosition: z
+    .object({
+      x: z.number(),
+      y: z.number(),
+    })
+    .optional()
+    .describe('Coordinates where pointer was shown'),
+})
+
+type CaptureScreenshotPointerInput = z.infer<
+  typeof CaptureScreenshotPointerInputSchema
+>
+type CaptureScreenshotPointerOutput = z.infer<
+  typeof CaptureScreenshotPointerOutputSchema
+>
+
+/**
+ * CaptureScreenshotPointerAction - Show pointer over element and capture screenshot
+ *
+ * Shows a visual pointer overlay at the center of the specified element,
+ * then captures a screenshot with the pointer visible.
+ *
+ * Prerequisites:
+ * - Must call getInteractiveSnapshot first to populate the cache
+ * - NodeId must exist in the cached snapshot
+ *
+ * Usage:
+ * 1. Get snapshot to find elements and populate cache
+ * 2. Call captureScreenshotPointer with tabId and nodeId
+ * 3. Returns screenshot with pointer overlay visible
+ *
+ * Used by: Visual debugging, automation demos, step-by-step captures
+ */
+export class CaptureScreenshotPointerAction extends ActionHandler<
+  CaptureScreenshotPointerInput,
+  CaptureScreenshotPointerOutput
+> {
+  readonly inputSchema = CaptureScreenshotPointerInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(
+    input: CaptureScreenshotPointerInput,
+  ): Promise<CaptureScreenshotPointerOutput> {
+    const { tabId, nodeId, size, pointerLabel } = input
+
+    // Get element rect from cache
+    const rect = SnapshotCache.getNodeRect(tabId, nodeId)
+
+    let pointerPosition: { x: number; y: number } | undefined
+
+    if (rect) {
+      // Calculate center coordinates
+      const { x, y } = PointerOverlay.getCenterCoordinates(rect)
+      pointerPosition = { x, y }
+
+      // Show pointer
+      await PointerOverlay.showPointer(tabId, x, y, pointerLabel)
+
+      logger.debug(
+        `[CaptureScreenshotPointerAction] Showed pointer at (${x}, ${y}) for node ${nodeId}`,
+      )
+    } else {
+      logger.warn(
+        `[CaptureScreenshotPointerAction] No cached rect for node ${nodeId} in tab ${tabId}. Capturing without pointer.`,
+      )
+    }
+
+    // Small delay to ensure pointer is rendered
+    await this.delay(100)
+
+    // Capture screenshot with pointer visible
+    const dataUrl = await this.browserOSAdapter.captureScreenshot(
+      tabId,
+      size as ScreenshotSizeKey | undefined,
+      false, // Don't show highlights, we have the pointer
+    )
+
+    return {
+      dataUrl,
+      pointerPosition,
+    }
+  }
+
+  private delay(ms: number): Promise<void> {
+    return new Promise((resolve) => setTimeout(resolve, ms))
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/ClearAction.ts

@@ -0,0 +1,38 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+const ClearInputSchema = z.object({
+  tabId: z.number().describe('The tab ID containing the element'),
+  nodeId: z
+    .number()
+    .int()
+    .positive()
+    .describe('The nodeId from interactive snapshot'),
+})
+
+type ClearInput = z.infer<typeof ClearInputSchema>
+interface ClearOutput {
+  success: boolean
+}
+
+/**
+ * ClearAction - Clear text from an input element
+ *
+ * Clears all text from an input field or textarea.
+ * Used before inputText or to reset form fields.
+ */
+export class ClearAction extends ActionHandler<ClearInput, ClearOutput> {
+  readonly inputSchema = ClearInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(input: ClearInput): Promise<ClearOutput> {
+    await this.browserOSAdapter.clear(input.tabId, input.nodeId)
+    return { success: true }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/ClickAction.ts

@@ -0,0 +1,62 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { PointerOverlay } from '@/utils/PointerOverlay'
+import { SnapshotCache } from '@/utils/SnapshotCache'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const ClickInputSchema = z.object({
+  tabId: z.number().describe('The tab ID containing the element'),
+  nodeId: z
+    .number()
+    .int()
+    .positive()
+    .describe('The nodeId from interactive snapshot'),
+})
+
+// Output schema
+const ClickOutputSchema = z.object({
+  success: z.boolean().describe('Whether the click succeeded'),
+})
+
+type ClickInput = z.infer<typeof ClickInputSchema>
+type ClickOutput = z.infer<typeof ClickOutputSchema>
+
+/**
+ * ClickAction - Click an element by its nodeId
+ *
+ * This action clicks an interactive element identified by its nodeId from getInteractiveSnapshot.
+ *
+ * Prerequisites:
+ * - Must call getInteractiveSnapshot first to get valid nodeIds
+ * - NodeIds are valid only for the current page state
+ * - NodeIds are invalidated on page navigation
+ *
+ * Usage:
+ * 1. Get snapshot to find clickable elements
+ * 2. Choose element by nodeId
+ * 3. Call click with tabId and nodeId
+ *
+ * Used by: ClickTool, all automation workflows
+ */
+export class ClickAction extends ActionHandler<ClickInput, ClickOutput> {
+  readonly inputSchema = ClickInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(input: ClickInput): Promise<ClickOutput> {
+    // Show pointer overlay before click
+    const rect = SnapshotCache.getNodeRect(input.tabId, input.nodeId)
+    if (rect) {
+      const { x, y } = PointerOverlay.getCenterCoordinates(rect)
+      await PointerOverlay.showPointerAndWait(input.tabId, x, y, 'Click')
+    }
+
+    await this.browserOSAdapter.click(input.tabId, input.nodeId)
+    return { success: true }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/ClickCoordinatesAction.ts

@@ -0,0 +1,69 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { getBrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { PointerOverlay } from '@/utils/PointerOverlay'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema for clickCoordinates action
+const ClickCoordinatesInputSchema = z.object({
+  tabId: z.number().int().positive().describe('Tab ID to click in'),
+  x: z.number().int().nonnegative().describe('X coordinate in viewport pixels'),
+  y: z.number().int().nonnegative().describe('Y coordinate in viewport pixels'),
+})
+
+type ClickCoordinatesInput = z.infer<typeof ClickCoordinatesInputSchema>
+
+// Output confirms the click
+export interface ClickCoordinatesOutput {
+  success: boolean
+  message: string
+  coordinates: {
+    x: number
+    y: number
+  }
+}
+
+/**
+ * ClickCoordinatesAction - Click at specific viewport coordinates
+ *
+ * Performs a click at the specified (x, y) coordinates in the viewport.
+ * Coordinates are in pixels relative to the top-left of the visible viewport (0, 0).
+ *
+ * Useful when:
+ * - Elements don't have accessible node IDs
+ * - Working with canvas or interactive graphics
+ * - Vision-based automation (e.g., AI identifies coordinates from screenshots)
+ *
+ * Example payload:
+ * {
+ *   "tabId": 123,
+ *   "x": 500,
+ *   "y": 300
+ * }
+ */
+export class ClickCoordinatesAction extends ActionHandler<
+  ClickCoordinatesInput,
+  ClickCoordinatesOutput
+> {
+  readonly inputSchema = ClickCoordinatesInputSchema
+  private browserOS = getBrowserOSAdapter()
+
+  async execute(input: ClickCoordinatesInput): Promise<ClickCoordinatesOutput> {
+    const { tabId, x, y } = input
+
+    // Show pointer overlay before click
+    await PointerOverlay.showPointerAndWait(tabId, x, y, 'Click')
+
+    await this.browserOS.clickCoordinates(tabId, x, y)
+
+    return {
+      success: true,
+      message: `Successfully clicked at coordinates (${x}, ${y}) in tab ${tabId}`,
+      coordinates: { x, y },
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/CloseWindowAction.ts

@@ -0,0 +1,38 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { CHROME_API_TIMEOUTS, withTimeout } from '@/utils/timeout'
+import { ActionHandler } from '../ActionHandler'
+
+const CloseWindowInputSchema = z.object({
+  windowId: z.number().int().positive().describe('ID of the window to close'),
+})
+
+const CloseWindowOutputSchema = z.object({
+  success: z.boolean().describe('Whether the window was successfully closed'),
+})
+
+type CloseWindowInput = z.infer<typeof CloseWindowInputSchema>
+type CloseWindowOutput = z.infer<typeof CloseWindowOutputSchema>
+
+export class CloseWindowAction extends ActionHandler<
+  CloseWindowInput,
+  CloseWindowOutput
+> {
+  readonly inputSchema = CloseWindowInputSchema
+
+  async execute(input: CloseWindowInput): Promise<CloseWindowOutput> {
+    await withTimeout(
+      chrome.windows.remove(input.windowId),
+      CHROME_API_TIMEOUTS.CHROME_API,
+      'chrome.windows.remove',
+    )
+
+    return {
+      success: true,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/CreateWindowAction.ts

@@ -0,0 +1,73 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { CHROME_API_TIMEOUTS, withTimeout } from '@/utils/timeout'
+import { ActionHandler } from '../ActionHandler'
+
+const CreateWindowInputSchema = z.object({
+  url: z
+    .string()
+    .optional()
+    .default('about:blank')
+    .describe('URL to open in the new window'),
+  incognito: z
+    .boolean()
+    .optional()
+    .default(false)
+    .describe('Create an incognito window'),
+  focused: z
+    .boolean()
+    .optional()
+    .default(true)
+    .describe('Whether to focus the new window'),
+})
+
+const CreateWindowOutputSchema = z.object({
+  windowId: z.number().describe('ID of the newly created window'),
+  tabId: z.number().describe('ID of the first tab in the new window'),
+})
+
+type CreateWindowInput = z.infer<typeof CreateWindowInputSchema>
+type CreateWindowOutput = z.infer<typeof CreateWindowOutputSchema>
+
+export class CreateWindowAction extends ActionHandler<
+  CreateWindowInput,
+  CreateWindowOutput
+> {
+  readonly inputSchema = CreateWindowInputSchema
+
+  async execute(input: CreateWindowInput): Promise<CreateWindowOutput> {
+    const createData: chrome.windows.CreateData = {
+      url: input.url,
+      focused: input.focused,
+      incognito: input.incognito,
+    }
+
+    const createdWindow = await withTimeout(
+      chrome.windows.create(createData),
+      CHROME_API_TIMEOUTS.CHROME_API,
+      'chrome.windows.create',
+    )
+
+    if (!createdWindow) {
+      throw new Error('Failed to create window')
+    }
+
+    if (createdWindow.id === undefined) {
+      throw new Error('Created window has no ID')
+    }
+
+    const tabId = createdWindow.tabs?.[0]?.id
+    if (tabId === undefined) {
+      throw new Error('Created window has no tab')
+    }
+
+    return {
+      windowId: createdWindow.id,
+      tabId,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/ExecuteJavaScriptAction.ts

@@ -0,0 +1,64 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const ExecuteJavaScriptInputSchema = z.object({
+  tabId: z.number().describe('The tab ID to execute code in'),
+  code: z.string().describe('JavaScript code to execute'),
+})
+
+// Output schema
+const ExecuteJavaScriptOutputSchema = z.object({
+  result: z.any().describe('The result of the code execution'),
+})
+
+type ExecuteJavaScriptInput = z.infer<typeof ExecuteJavaScriptInputSchema>
+type ExecuteJavaScriptOutput = z.infer<typeof ExecuteJavaScriptOutputSchema>
+
+/**
+ * ExecuteJavaScriptAction - Execute JavaScript code in page context
+ *
+ * Executes arbitrary JavaScript code in the page and returns the result.
+ *
+ * Input:
+ * - tabId: Tab ID to execute code in
+ * - code: JavaScript code as string
+ *
+ * Output:
+ * - result: The return value of the executed code
+ *
+ * Usage:
+ * - Extract data from page: "document.title"
+ * - Manipulate DOM: "document.body.style.background = 'red'"
+ * - Get element values: "document.querySelector('#email').value"
+ *
+ * Example:
+ * {
+ *   "tabId": 123,
+ *   "code": "document.title"
+ * }
+ * // Returns: { result: "Google" }
+ */
+export class ExecuteJavaScriptAction extends ActionHandler<
+  ExecuteJavaScriptInput,
+  ExecuteJavaScriptOutput
+> {
+  readonly inputSchema = ExecuteJavaScriptInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(
+    input: ExecuteJavaScriptInput,
+  ): Promise<ExecuteJavaScriptOutput> {
+    const result = await this.browserOSAdapter.executeJavaScript(
+      input.tabId,
+      input.code,
+    )
+    return { result }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/GetAccessibilityTreeAction.ts

@@ -0,0 +1,53 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+const GetAccessibilityTreeInputSchema = z.object({
+  tabId: z
+    .number()
+    .int()
+    .positive()
+    .describe('Tab ID to get accessibility tree from'),
+})
+
+type GetAccessibilityTreeInput = z.infer<typeof GetAccessibilityTreeInputSchema>
+export type GetAccessibilityTreeOutput = chrome.browserOS.AccessibilityTree
+
+/**
+ * GetAccessibilityTreeAction - Get accessibility tree for a tab
+ *
+ * Returns the full accessibility tree structure containing:
+ * - rootId: The root node ID
+ * - nodes: Map of node IDs to accessibility nodes
+ *
+ * Each node contains:
+ * - nodeId: Unique node identifier
+ * - role: Accessibility role (e.g., 'staticText', 'heading', 'button')
+ * - name: Text content or label
+ * - childIds: Array of child node IDs
+ *
+ * Example payload:
+ * {
+ *   "tabId": 123
+ * }
+ */
+export class GetAccessibilityTreeAction extends ActionHandler<
+  GetAccessibilityTreeInput,
+  GetAccessibilityTreeOutput
+> {
+  readonly inputSchema = GetAccessibilityTreeInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(
+    input: GetAccessibilityTreeInput,
+  ): Promise<GetAccessibilityTreeOutput> {
+    const { tabId } = input
+    const tree = await this.browserOSAdapter.getAccessibilityTree(tabId)
+    return tree
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/GetInteractiveSnapshotAction.ts

@@ -0,0 +1,71 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import type {
+  InteractiveSnapshot,
+  InteractiveSnapshotOptions,
+} from '@/adapters/BrowserOSAdapter'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { SnapshotCache } from '@/utils/SnapshotCache'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const GetInteractiveSnapshotInputSchema = z.object({
+  tabId: z.number().describe('The tab ID to get snapshot from'),
+  options: z
+    .object({
+      includeHidden: z
+        .boolean()
+        .optional()
+        .default(false)
+        .describe('Include hidden elements (default: false)'),
+    })
+    .optional()
+    .describe('Optional snapshot options'),
+})
+
+type GetInteractiveSnapshotInput = z.infer<
+  typeof GetInteractiveSnapshotInputSchema
+>
+
+/**
+ * GetInteractiveSnapshotAction - Get interactive elements from the page
+ *
+ * This is THE MOST CRITICAL action - it returns all interactive elements
+ * with their nodeIds, which are needed by click, inputText, clear, and scrollToNode actions.
+ *
+ * Returns:
+ * - elements: Array of interactive nodes with nodeIds
+ * - hierarchicalStructure: String representation of page structure
+ *
+ * Each element contains:
+ * - nodeId: Sequential integer ID (1, 2, 3...)
+ * - type: 'clickable' | 'typeable' | 'selectable'
+ * - name: Element text/label
+ * - attributes: Element properties (html-tag, role, etc.)
+ * - rect: Bounding box coordinates
+ */
+export class GetInteractiveSnapshotAction extends ActionHandler<
+  GetInteractiveSnapshotInput,
+  InteractiveSnapshot
+> {
+  readonly inputSchema = GetInteractiveSnapshotInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(
+    input: GetInteractiveSnapshotInput,
+  ): Promise<InteractiveSnapshot> {
+    const snapshot = await this.browserOSAdapter.getInteractiveSnapshot(
+      input.tabId,
+      input.options as InteractiveSnapshotOptions | undefined,
+    )
+
+    // Cache snapshot for pointer overlay lookup
+    SnapshotCache.set(input.tabId, snapshot)
+
+    return snapshot
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/GetPageLoadStatusAction.ts

@@ -0,0 +1,69 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import {
+  BrowserOSAdapter,
+  type PageLoadStatus,
+} from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema for getPageLoadStatus action
+const GetPageLoadStatusInputSchema = z.object({
+  tabId: z
+    .number()
+    .int()
+    .positive()
+    .describe('Tab ID to check page load status'),
+})
+
+type GetPageLoadStatusInput = z.infer<typeof GetPageLoadStatusInputSchema>
+
+// Output includes page load status details
+export interface GetPageLoadStatusOutput {
+  tabId: number
+  isResourcesLoading: boolean
+  isDOMContentLoaded: boolean
+  isPageComplete: boolean
+}
+
+/**
+ * GetPageLoadStatusAction - Get page loading status for a tab
+ *
+ * Returns the current page load status including:
+ * - isResourcesLoading: Whether resources (images, scripts, etc.) are still loading
+ * - isDOMContentLoaded: Whether the DOM is fully parsed and ready
+ * - isPageComplete: Whether the page has completely finished loading
+ *
+ * Useful for waiting for pages to load before taking actions.
+ *
+ * Example payload:
+ * {
+ *   "tabId": 123
+ * }
+ */
+export class GetPageLoadStatusAction extends ActionHandler<
+  GetPageLoadStatusInput,
+  GetPageLoadStatusOutput
+> {
+  readonly inputSchema = GetPageLoadStatusInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(
+    input: GetPageLoadStatusInput,
+  ): Promise<GetPageLoadStatusOutput> {
+    const { tabId } = input
+
+    const status: PageLoadStatus =
+      await this.browserOSAdapter.getPageLoadStatus(tabId)
+
+    return {
+      tabId,
+      isResourcesLoading: status.isResourcesLoading,
+      isDOMContentLoaded: status.isDOMContentLoaded,
+      isPageComplete: status.isPageComplete,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/GetSnapshotAction.ts

@@ -0,0 +1,74 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter, type Snapshot } from '@/adapters/BrowserOSAdapter'
+import { logger } from '@/utils/logger'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema for getSnapshot action
+const GetSnapshotInputSchema = z.object({
+  tabId: z.number().int().positive().describe('Tab ID to get snapshot from'),
+  type: z
+    .enum(['text', 'links'])
+    .default('text')
+    .describe('Type of snapshot: text or links'),
+  options: z
+    .object({
+      context: z.enum(['visible', 'full']).optional(),
+      includeSections: z
+        .array(
+          z.enum([
+            'main',
+            'navigation',
+            'footer',
+            'header',
+            'article',
+            'aside',
+          ]),
+        )
+        .optional(),
+    })
+    .optional()
+    .describe('Optional snapshot configuration'),
+})
+
+type GetSnapshotInput = z.infer<typeof GetSnapshotInputSchema>
+
+// Output is the full snapshot structure
+export type GetSnapshotOutput = Snapshot
+
+/**
+ * GetSnapshotAction - Extract page content snapshot
+ *
+ * Extracts structured content from the page including:
+ * - Headings (with levels)
+ * - Text content
+ * - Links (with URLs)
+ *
+ * Returns items in document order with type information.
+ *
+ * Example payload:
+ * {
+ *   "tabId": 123,
+ *   "type": "text"
+ * }
+ */
+export class GetSnapshotAction extends ActionHandler<
+  GetSnapshotInput,
+  GetSnapshotOutput
+> {
+  readonly inputSchema = GetSnapshotInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(input: GetSnapshotInput): Promise<GetSnapshotOutput> {
+    const { tabId, type } = input
+    logger.info(
+      `[GetSnapshotAction] Getting snapshot for tab ${tabId} with type ${type}`,
+    )
+    const snapshot = await this.browserOSAdapter.getSnapshot(tabId, type)
+    return snapshot
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/InputTextAction.ts

@@ -0,0 +1,75 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { PointerOverlay } from '@/utils/PointerOverlay'
+import { SnapshotCache } from '@/utils/SnapshotCache'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const InputTextInputSchema = z.object({
+  tabId: z.number().describe('The tab ID containing the element'),
+  nodeId: z
+    .number()
+    .int()
+    .positive()
+    .describe('The nodeId from interactive snapshot'),
+  text: z.string().describe('Text to type into the element'),
+})
+
+// Output schema
+const InputTextOutputSchema = z.object({
+  success: z.boolean().describe('Whether the input succeeded'),
+})
+
+type InputTextInput = z.infer<typeof InputTextInputSchema>
+type InputTextOutput = z.infer<typeof InputTextOutputSchema>
+
+/**
+ * InputTextAction - Type text into an element by its nodeId
+ *
+ * This action types text into an input field or textarea identified by its nodeId.
+ *
+ * Prerequisites:
+ * - Must call getInteractiveSnapshot first to get valid nodeIds
+ * - Element must be typeable (type: 'typeable' in snapshot)
+ * - NodeIds are valid only for the current page state
+ *
+ * Behavior:
+ * - Automatically clears existing text before typing (handled by adapter)
+ * - Types the full text string
+ * - Triggers input/change events
+ *
+ * Usage:
+ * 1. Get snapshot to find typeable elements
+ * 2. Choose input field by nodeId
+ * 3. Call inputText with tabId, nodeId, and text
+ *
+ * Used by: TypeTool, form automation workflows
+ */
+export class InputTextAction extends ActionHandler<
+  InputTextInput,
+  InputTextOutput
+> {
+  readonly inputSchema = InputTextInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(input: InputTextInput): Promise<InputTextOutput> {
+    // Show pointer overlay before typing
+    const rect = SnapshotCache.getNodeRect(input.tabId, input.nodeId)
+    if (rect) {
+      const { x, y } = PointerOverlay.getLeftCenterCoordinates(rect)
+      const textPreview =
+        input.text.length > 20
+          ? `Type: ${input.text.substring(0, 20)}...`
+          : `Type: ${input.text}`
+      await PointerOverlay.showPointerAndWait(input.tabId, x, y, textPreview)
+    }
+
+    await this.browserOSAdapter.inputText(input.tabId, input.nodeId, input.text)
+    return { success: true }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/ScrollDownAction.ts

@@ -0,0 +1,54 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const ScrollDownInputSchema = z.object({
+  tabId: z.number().describe('The tab ID to scroll'),
+})
+
+// Output schema
+const ScrollDownOutputSchema = z.object({
+  success: z.boolean().describe('Whether the scroll succeeded'),
+})
+
+type ScrollDownInput = z.infer<typeof ScrollDownInputSchema>
+type ScrollDownOutput = z.infer<typeof ScrollDownOutputSchema>
+
+/**
+ * ScrollDownAction - Scroll page down
+ *
+ * Scrolls the page down by one viewport height using PageDown key.
+ * This approach is more reliable than the direct scrollDown API.
+ *
+ * Input:
+ * - tabId: Tab ID to scroll
+ *
+ * Output:
+ * - success: true if scroll succeeded
+ *
+ * Usage:
+ * Used for scrolling through long pages to view content below the fold.
+ */
+export class ScrollDownAction extends ActionHandler<
+  ScrollDownInput,
+  ScrollDownOutput
+> {
+  readonly inputSchema = ScrollDownInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(input: ScrollDownInput): Promise<ScrollDownOutput> {
+    // Use sendKeys with PageDown instead of scrollDown API (more reliable)
+    await this.browserOSAdapter.sendKeys(input.tabId, 'PageDown')
+
+    // Add small delay for scroll to complete
+    await new Promise((resolve) => setTimeout(resolve, 100))
+
+    return { success: true }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/ScrollToNodeAction.ts

@@ -0,0 +1,42 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+const ScrollToNodeInputSchema = z.object({
+  tabId: z.number().describe('The tab ID containing the element'),
+  nodeId: z.number().int().positive().describe('The nodeId to scroll to'),
+})
+
+type ScrollToNodeInput = z.infer<typeof ScrollToNodeInputSchema>
+interface ScrollToNodeOutput {
+  scrolled: boolean
+}
+
+/**
+ * ScrollToNodeAction - Scroll an element into view
+ *
+ * Scrolls the page so that the specified element is visible in the viewport.
+ * Returns whether scrolling actually occurred.
+ *
+ * Used by: Click/Type tools to ensure element is visible before interaction
+ */
+export class ScrollToNodeAction extends ActionHandler<
+  ScrollToNodeInput,
+  ScrollToNodeOutput
+> {
+  readonly inputSchema = ScrollToNodeInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(input: ScrollToNodeInput): Promise<ScrollToNodeOutput> {
+    const scrolled = await this.browserOSAdapter.scrollToNode(
+      input.tabId,
+      input.nodeId,
+    )
+    return { scrolled }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/ScrollUpAction.ts

@@ -0,0 +1,54 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { BrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema
+const ScrollUpInputSchema = z.object({
+  tabId: z.number().describe('The tab ID to scroll'),
+})
+
+// Output schema
+const ScrollUpOutputSchema = z.object({
+  success: z.boolean().describe('Whether the scroll succeeded'),
+})
+
+type ScrollUpInput = z.infer<typeof ScrollUpInputSchema>
+type ScrollUpOutput = z.infer<typeof ScrollUpOutputSchema>
+
+/**
+ * ScrollUpAction - Scroll page up
+ *
+ * Scrolls the page up by one viewport height using PageUp key.
+ * This approach is more reliable than the direct scrollUp API.
+ *
+ * Input:
+ * - tabId: Tab ID to scroll
+ *
+ * Output:
+ * - success: true if scroll succeeded
+ *
+ * Usage:
+ * Used for scrolling back up through long pages.
+ */
+export class ScrollUpAction extends ActionHandler<
+  ScrollUpInput,
+  ScrollUpOutput
+> {
+  readonly inputSchema = ScrollUpInputSchema
+  private browserOSAdapter = BrowserOSAdapter.getInstance()
+
+  async execute(input: ScrollUpInput): Promise<ScrollUpOutput> {
+    // Use sendKeys with PageUp instead of scrollUp API (more reliable)
+    await this.browserOSAdapter.sendKeys(input.tabId, 'PageUp')
+
+    // Add small delay for scroll to complete
+    await new Promise((resolve) => setTimeout(resolve, 100))
+
+    return { success: true }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/SendKeysAction.ts

@@ -0,0 +1,69 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { getBrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema for sendKeys action
+const SendKeysInputSchema = z.object({
+  tabId: z.number().int().positive().describe('Tab ID to send keys to'),
+  key: z
+    .enum([
+      'Enter',
+      'Delete',
+      'Backspace',
+      'Tab',
+      'Escape',
+      'ArrowUp',
+      'ArrowDown',
+      'ArrowLeft',
+      'ArrowRight',
+      'Home',
+      'End',
+      'PageUp',
+      'PageDown',
+    ])
+    .describe('Keyboard key to send'),
+})
+
+type SendKeysInput = z.infer<typeof SendKeysInputSchema>
+
+// Output is just success (void result)
+export interface SendKeysOutput {
+  success: boolean
+  message: string
+}
+
+/**
+ * SendKeysAction - Send keyboard keys to a tab
+ *
+ * Sends special keyboard keys (Enter, Escape, arrows, etc.) to the specified tab.
+ * Useful for navigation, form submission, closing dialogs, etc.
+ *
+ * Example payload:
+ * {
+ *   "tabId": 123,
+ *   "key": "Enter"
+ * }
+ */
+export class SendKeysAction extends ActionHandler<
+  SendKeysInput,
+  SendKeysOutput
+> {
+  readonly inputSchema = SendKeysInputSchema
+  private browserOS = getBrowserOSAdapter()
+
+  async execute(input: SendKeysInput): Promise<SendKeysOutput> {
+    const { tabId, key } = input
+
+    await this.browserOS.sendKeys(tabId, key as chrome.browserOS.Key)
+
+    return {
+      success: true,
+      message: `Successfully sent "${key}" to tab ${tabId}`,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/browser/TypeAtCoordinatesAction.ts

@@ -0,0 +1,81 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+import { getBrowserOSAdapter } from '@/adapters/BrowserOSAdapter'
+import { PointerOverlay } from '@/utils/PointerOverlay'
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema for typeAtCoordinates action
+const TypeAtCoordinatesInputSchema = z.object({
+  tabId: z.number().int().positive().describe('Tab ID to type in'),
+  x: z.number().int().nonnegative().describe('X coordinate in viewport pixels'),
+  y: z.number().int().nonnegative().describe('Y coordinate in viewport pixels'),
+  text: z.string().min(1).describe('Text to type at the location'),
+})
+
+type TypeAtCoordinatesInput = z.infer<typeof TypeAtCoordinatesInputSchema>
+
+// Output confirms the typing
+export interface TypeAtCoordinatesOutput {
+  success: boolean
+  message: string
+  coordinates: {
+    x: number
+    y: number
+  }
+  textLength: number
+}
+
+/**
+ * TypeAtCoordinatesAction - Type text at specific viewport coordinates
+ *
+ * Clicks at the specified (x, y) coordinates and types the provided text.
+ * Coordinates are in pixels relative to the top-left of the visible viewport (0, 0).
+ *
+ * The action will:
+ * 1. Click at the coordinates to focus the element
+ * 2. Type the specified text
+ *
+ * Useful when:
+ * - Input fields don't have accessible node IDs
+ * - Working with complex forms or canvas-based inputs
+ * - Vision-based automation (e.g., AI identifies input coordinates from screenshots)
+ *
+ * Example payload:
+ * {
+ *   "tabId": 123,
+ *   "x": 500,
+ *   "y": 300,
+ *   "text": "Hello World"
+ * }
+ */
+export class TypeAtCoordinatesAction extends ActionHandler<
+  TypeAtCoordinatesInput,
+  TypeAtCoordinatesOutput
+> {
+  readonly inputSchema = TypeAtCoordinatesInputSchema
+  private browserOS = getBrowserOSAdapter()
+
+  async execute(
+    input: TypeAtCoordinatesInput,
+  ): Promise<TypeAtCoordinatesOutput> {
+    const { tabId, x, y, text } = input
+
+    // Show pointer overlay before typing
+    const textPreview =
+      text.length > 20 ? `Type: ${text.substring(0, 20)}...` : `Type: ${text}`
+    await PointerOverlay.showPointerAndWait(tabId, x, y, textPreview)
+
+    await this.browserOS.typeAtCoordinates(tabId, x, y, text)
+
+    return {
+      success: true,
+      message: `Successfully typed "${text.substring(0, 50)}${text.length > 50 ? '...' : ''}" at coordinates (${x}, ${y}) in tab ${tabId}`,
+      coordinates: { x, y },
+      textLength: text.length,
+    }
+  }
+}

packages/browseros-agent/apps/controller-ext/src/actions/diagnostics/CheckBrowserOSAction.ts

@@ -0,0 +1,90 @@
+/**
+ * @license
+ * Copyright 2025 BrowserOS
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ */
+import { z } from 'zod'
+
+import { ActionHandler } from '../ActionHandler'
+
+// Input schema - no input needed
+const CheckBrowserOSInputSchema = z.any()
+
+// Output schema
+const CheckBrowserOSOutputSchema = z.object({
+  available: z.boolean(),
+  apis: z.array(z.string()).optional(),
+  error: z.string().optional(),
+})
+
+type CheckBrowserOSInput = z.infer<typeof CheckBrowserOSInputSchema>
+type CheckBrowserOSOutput = z.infer<typeof CheckBrowserOSOutputSchema>
+
+/**
+ * CheckBrowserOSAction - Diagnostic action to check if chrome.browserOS is available
+ *
+ * This action checks:
+ * 1. Whether chrome.browserOS namespace exists
+ * 2. What APIs are available in the namespace
+ * 3. Returns detailed diagnostic information
+ */
+export class CheckBrowserOSAction extends ActionHandler<
+  CheckBrowserOSInput,
+  CheckBrowserOSOutput
+> {
+  readonly inputSchema = CheckBrowserOSInputSchema
+
+  async execute(_input: CheckBrowserOSInput): Promise<CheckBrowserOSOutput> {
+    try {
+      console.log('[CheckBrowserOSAction] Starting diagnostic...')
+      console.log('[CheckBrowserOSAction] typeof chrome:', typeof chrome)
+      console.log('[CheckBrowserOSAction] chrome exists:', chrome !== undefined)
+
+      // Check if chrome.browserOS exists
+      const browserOSExists = typeof chrome.browserOS !== 'undefined'
+      console.log(
+        '[CheckBrowserOSAction] typeof chrome.browserOS:',
+        typeof chrome.browserOS,
+      )
+      console.log('[CheckBrowserOSAction] browserOSExists:', browserOSExists)
+
+      if (!browserOSExists) {
+        console.log('[CheckBrowserOSAction] chrome.browserOS is NOT available')
+        return {
+          available: false,
+          error:
+            'chrome.browserOS is undefined - not running in BrowserOS Chrome',
+        }
+      }
+
+      // Get available APIs
+      const apis: string[] = []
+      const browserOS = chrome.browserOS as Record<string, unknown>
+
+      for (const key in browserOS) {
+        if (typeof browserOS[key] === 'function') {
+          apis.push(key)
+        }
+      }
+
+      console.log('[CheckBrowserOSAction] Found APIs:', apis)
+
+      return {
+        available: true,
+        apis: apis.sort(),
+      }
+    } catch (error) {
+      console.error('[CheckBrowserOSAction] Error during diagnostic:', error)
+      const errorMsg =
+        error instanceof Error
+          ? error.message
+          : error
+            ? String(error)
+            : 'Unknown error'
+      return {
+        available: false,
+        error: errorMsg,
+      }
+    }
+  }
+}